The backdrop to this week’s Budget is gloomy. Britain’s high streets remain a grim collage of betting shops and vape stores. Public services that were outdated a decade ago now groan under their own bloat. Prices rise and pay packets stagnate. Young people go without work or training, opportunity drying up as the cost of creating it climbs. Two-thirds of British adults think today’s children will be worse off than their parents. In such a country, narratives of doomerism and decline are all too easy to reach for.

But I for one am optimistic about Britain’s future. Our country has the capability not just to survive in our radically changing world, but to thrive in it. It has endured seven shocks in 18 years — two of which were self-inflicted — and yet it is still standing. Britain’s economy is clearly resilient. The question now facing John Healey is how to grow it.

The answer is AI. Love it or hate it, it doesn’t matter: because AI is “the thing” and trillions of dollars are being poured into it for good reason. It’s a transformative technology that is here to stay, whether or not the bubble surrounding it bursts, just as the railways and the internet outlived the bubbles that built them.

Clearly, AI presents profound challenges. It will change the labor market, destroying jobs and creating new ones. It threatens the climate but could also heal it. And it throws open the question of sovereignty: which values will govern the world when a handful of American and Chinese firms own the models remaking it?

So it is not AI in and of itself that fills me with optimism, but Britain’s ability to shape it. The technology will create and it will destroy. But of all the countries it will transform, few are better placed to harness it to bring about progress and prosperity than Britain. The Chancellor’s Budget this week must recognize this. It should be a watershed moment that ends the old political imprisonment of tax and spend — now busted, as taxes reach postwar highs, and spending eye-watering levels — and ultimately secures Britain’s place in our new AI world. 

Just think about how AI could transform the NHS, which already holds the cradle-to-grave records of almost 70 million people, built over decades. This data is as valuable as any tax rise or spending cut. And, powered by AI, it is a stepping stone to creating a truly preventative health service. Yes, the politics of using it are hard (google Palantir and you’ll understand why) but we shouldn’t cower from making the argument for a data-driven, cheaper, better NHS. Health spending has grown by 3.6% every year, on average, since the NHS was founded — which is much faster than the economy that pays for it. Britain is fatter and sicker than its neighbours, with 2.5 million people out of work due to ill health. This is an economic and moral crisis, and yet we are still debating spending levels, watching as game-changing weight-loss drugs improve the lives of those who can afford them and struggle to reach those who need them. 

A preventative healthcare system, powered by AI and able to capitalize on the latest scientific discoveries, will be able to stop a heart attack or a stroke before it happens. It should be central to the Chancellor’s Budget, and offer a template for wider public-service reform, with every government department seeking to accelerate AI-enabled transformation.

A similar approach could be taken to the Prime Minister’s proposed national care service. There is a risk this becomes yet another burden on Britain’s finances. So it would be better to look at the root of the problem — dementia — which affects around four-in-five care-home residents and costs the country £42 billion a year. Just 1.4% of NHS spending on dementia goes to diagnosis and treatment — this proportion should be far higher. Britain’s science and technology sector is powering a revolution in dementia treatment. The first drugs that slow Alzheimer’s already exist and the next generation can be cheaper and better. DeepMind, which was founded in London, cracked protein folding, contributing to the transformation of medicine development; its spin-out, Isomorphic, now designs medicines with its tools. 

No other country apart from the US is better placed to seize this moment. AI is an expertise accelerant and Britain has an economy built on expertise. We could train AI on the best legal system in the world — English law governs around 40% of global business and financial transactions — or on the Premier League’s data, the BBC’s archive, the City’s markets or the research of our top-tier universities. It’s a recipe for wealth-creating unicorns. 

Suggested reading

The Budget is Britain’s tribal ritual

By Terry Eagleton

We are on our way already. British start-ups raised $23 billion in the first nine months of this year, more than a third of everything raised in Europe. And our manufacturing base, which we often bemoan as a shell of its former self, is now churning out cutting-edge equipment. Rolls-Royce builds the reactors for our submarines in Derby, while Broughton builds the wings for Airbus. These are rare, best-in-class products, and making them here is a boost to our sovereignty and power. These are products most countries can’t make and none can go without. 

This is why the Budget should encourage even more innovation, particularly in the defense industry. It should, for instance, lay the groundwork for manufacturing AI robots with arms and legs, which could be used for both civilian and military purposes. Or it could back apprenticeships at scale, to teach young people how to build the defenses our country and its allies need. This would be as good an answer to a million young people out of work and training as any benefit reform. But for this, the Chancellor would need to prioritize defense spending in the interest of national security. 

Meanwhile, Britain has quietly built leverage in the AI race. The world’s frontier models are tested at our AI Security Institute before they are released, putting Britain in the room where the rules are written. As Britain’s very able AI minister Kanishka Narayan told me recently: we build “capability tonight, convening tomorrow, leverage the day after”. 

The biggest barrier to this future? Us. We are our own worst enemies. Sure, a yawning deficit and astronomical debt don’t help, nor does the fact that the UK’s investment levels have been the lowest in the G7 in 24 of the last 30 years. But if we choose to embrace the AI revolution rather than continuing with the status quo of decline, we will have greater influence on our future. 

It’s not the easy choice. If the Chancellor does make this the Budget that harnesses AI’s potential, he will face a British system built to say no: a broken planning regime; a civil service too scared to act; pension funds that won’t back the country they operate in; a dysfunctional energy system pushing bills through the roof. 

He will need to fix the energy system first: we are paying almost £2 billion a year to switch wind farms off because the grid can’t carry their power to the places where it’s needed, leaving us with some of the most expensive electricity in the world. Investors make decisions worth billions on that fact. The approval of the Jackdaw gas field was a flicker of pragmatism, and Rosebank should follow. But clean power, rather than cheap power, remains the overriding goal of energy policy, even as the facts scream at us to change course. Only this month, Microsoft pulled out of a landmark deal with a data center in Essex because the site couldn’t access electricity before 2030. The goal of Cheap Power 2030 should therefore replace Clean Power 2030 on Wednesday.

If this Budget falls short of being the first step in reversing Britain’s decline, then doomerism will evolve from mood to habit. People will hesitate before starting a business or backing one. They will think twice about retraining or upskilling because they no longer believe a better future is possible. The temptation for political leaders may be to adopt easy retail politics, or to retreat into nostalgia, pointing to those better days gone by. But leadership of the highest order — which we need now and haven’t seen for some time — isn’t about trying to work out what voters want and giving it to them, but working out what the country needs, convincing voters it’s what they want, and taking them into the future. 

In this case, AI might just have saved the day. The flaws were discovered because cybersecurity experts used AI to search for weaknesses that humans had missed. Had, say, AI-assisted Russian cybercriminals got there first, it might have been far worse. The incident gives an insight into the double-edged nature of AI, which presents clear and present dangers — but also, new tools to combat those dangers. Note, however, that we are talking about the fairly unsexy business of cybercrime  and not, for example, sentient computers spontaneously deciding to eradicate humanity or farming our living souls in some sort of Matrix nightmare. And yet it is these scenarios, dreamed up by a small set of extravagantly funded AI doomers and spread by a credulous media that have occupied the world’s attention; and have gone into overdrive since the Hugging Face hack and Jacob Coxon’s resignation from Anthropic.

For “normies”, anxiety about a new emerging technology tends to center on things like jobs, education, or the economy. Cybersecurity experts tend to worry about tangible threats to existing systems. But AI doomerism is distinct. It describes people who believe that advanced AI poses a risk of human extinction — and are, generally, prepared to put a number on it. Coxon’s former colleague Evan Hubinger set his own “p(doom)” at “above 10%”; some go as high as 25% or even 50% .

Suggested reading

Why is Britain blocking AI?

By Tom Ough

Much has been written about the strange confluence of money, power, and quasi-theological belief systems that have given rise to this apocalyptic worldview. Still, what receives less attention is what this movement actually wants to happen in order to avert this admittedly unpleasant scenario — which is to say, the authoritarian and even totalitarian implication of the AI doomers’ policy proposals. Ideas of mass surveillance that would make Big Brother blush have almost imperceptibly crept from the fantastical thought experiments of Effective Altruists and the Rationalists and are being now turned into policy proposals by such figures as Bernie Sanders.

To understand the peculiar flavor of AI doomerism, a good place to start is a 2019 TED talk given by Oxford’s Nick Bostrom, one of the leading philosophers on existential risks, titled “How civilization could destroy itself—and 4 ways we could prevent it”. Bostrom opened his talk with a thought experiment. He asked his audience to imagine an urn containing various balls representing technological advances. There would be some white balls, representing beneficial discoveries. There would be some gray balls, representing discoveries that have some positive and some negative benefits. But there might be a few black balls: discoveries so profound, they would effectively make political life as we know it unsustainable. One example might be synthetic biology: imagine if your average supervillain really did gain the ability to create mutants! Another might be AGI.

Bostrom’s point was that once the ball is withdrawn from the urn, you can’t put it back. Therefore, the only viable technological safeguard would be to stop anyone from withdrawing any black balls in the first place. And the only real way of doing this, he mused, was: “Ubiquitous surveillance — everybody would be monitored all the time.” The interviewer’s thoughts turned immediately to dystopia: “This is Minority Report, essentially,” he said, referencing the 2002 Tom Cruise movie about a future police state that uses psychics to arrest criminals before they have a chance to commit their crimes. Bostrom duly admitted that there were “huge problems and risks” associated with the idea of mass surveillance and global governance. But perhaps, he mused, people would be more willing to wear multidirectional cameras and microphones if we called this “The ‘freedom tag’… or something like that”. Still, he suggested that “we seemed to be doomed anyway”, by which logic, extreme solutions may come to seem proportionate.

Over time, this “end justifies the means” mindset moved from thought experiment to policy proposal, with increasingly concrete ideas for controlling AI research and the people working on it. With the work of the Machine Intelligence Research Institute (MIRI), the case for totalizing enforcement became more explicit.

MIRI is a Berkeley-based nonprofit organization, founded more than two decades ago by the writer Eliezer Yudkowsky and backed by a remarkable concentration of money and influence. Its funders include Open Philanthropy, the philanthropic organization founded by Facebook co-founder Dustin Moskovitz and Cari Tuna; Vitalik Buterin, the co-founder of Ethereum (an open-source blockchain); and Jaan Tallinn, the Estonian programmer who helped build Skype and has since made existential risk one of the central objects of his philanthropy. Tallinn is a particularly important figure in this ecosystem: through the Survival and Flourishing Fund, which he funds alongside other philanthropists, he has helped direct hundreds of millions of dollars towards organizations concerned with AI existential risks.

MIRI’s particular obsession is artificial superintelligence (ASI) — a hypothetical scenario in which computers develop vastly superior intelligence to humans on every vector. Its position on ASI is: “Everyone dies.” Yudkowsky therefore argues that the only reasonable response is a globally coordinated and collectively enforced moratorium on the development of ASI.

This may not worry you, since you (I assume) have no intention of trying to create ASI. But the wider problem is enforcement. If governments agree to halt this kind of research, how do they know companies or individuals are obeying? In June, MIRI’s Technical Governance Team published Verifying Restrictions on Frontier AI Research, which considers how such a prohibition could actually be policed. The paper catalogs 28 possible verification mechanisms, including intelligence gathering; international search warrants and inspections (of properties, computers, and files); polygraphs; monitoring what users type into AI systems and what the model produces while it is being used; sting operations; employee-monitoring software; and AI-assisted code review, which may extend to codebases, experiment logs, and research documents. It is a proposal for giving governments unprecedented oversight of private activity. Researchers would work under permanent clouds of suspicion, never knowing which of their experiments might be deemed illicit, which conversations were incriminating, which of their colleagues was an informant.

This really is a dystopian nightmare, and, unfortunately, not the dystopia the media is covering. Journalists have spent years asking readers to imagine the dictatorship of future machines while paying far less attention to the systems of human control being proposed to prevent it.

MIRI itself acknowledged that such mechanisms could “enable authoritarianism” and cautioned that some “might be undesirable.” And yet thanks to the well-funded ecosystem of organizations surrounding MIRI, these ideas have now migrated into public policy proposals.

On 23 September, Senator Bernie Sanders and Representative Greg Casar introduced the Ban Artificial Superintelligence Act, which would permanently prohibit the development and deployment of artificial superintelligence and immediately pause advanced AI development until a new federal regulator was established to enforce the bans and strict penalties. The bill would create a cabinet-level Department of Artificial Intelligence with powers to monitor frontier AI systems and oversee their destruction. Individuals who violated or circumvented the ban could face up to 20 years in prison; companies could face what the sponsors call the “corporate death penalty”. That is a remarkable amount of government power over a technology that doesn’t even exist.

“A new age of AI radicalism is upon us,” concluded Adam Thierer and Nicholas Thielman of the R Street Institute, describing the Sanders-Casar framework as one of “the most radical interventions ever proposed in any piece of proposed legislation in American history”. Their report warned that “if enacted, it would have extremely destructive and dangerous consequences for innovation, competition, economic growth, global competitiveness, national security, and freedom of speech.”

They label the proposals “AI Authoritarianism”, arguing that it represents an unprecedented expansion of government control over technology and potentially over information itself, sharply restricting what companies, scientists, and individuals are allowed to do. Even pro-compute governance groups such as GovAI have cautioned that such a regime could “also be used to infringe on civil liberties, perpetuate existing power structures, and entrench authoritarian regimes.” Indeed, it could reasonably be argued that its implications are totalitarian — since enforcement will necessarily extend into private intellectual activity itself, through the monitoring of researchers’ prompts, code, messages, and devices via a pervasive surveillance system designed to prevent even private non-compliance.

These ideas have also found an energetic British advocate in ControlAI, a non-profit founded by Andrea Miotti to campaign against the existential risk posed by superintelligent AI. Its policy proposal also includes a 20-year AI pause. Backed by figures including Tallinn, the organization has spent the past two years briefing politicians in Britain, America, and elsewhere. In September, its UK campaign achieved a milestone: Labour MP Alex Sobel introduced a bill, drafted by ControlAI, that would ban the development of superintelligent AI in Britain for 20 years and require the Government to pursue an international agreement banning it worldwide. ControlAI also drafted a separate AI “Kill-Switch” amendment, designed to grant the government emergency powers to shut down AI models and data centers. And its US arm also consulted with the offices of Bernie Sanders and Greg Casar on their Ban Artificial Superintelligence Act.

The moratorium is extremely important to the AI doomers. Writing in Time in 2023, Yudkowsky warned readers: “Be less scared of a shooting conflict between nations than of the moratorium being violated; be willing to destroy a rogue datacenter by airstrike.” Yudkowsky’s argument treats interstate military conflict, potentially involving nuclear-armed powers, as a lesser danger than his hypothetical risk of AI-induced catastrophe. In that framework, even conventional geopolitical dangers become subordinate to the threat he assigns to AI.

If it were up to Nick Bostrom and Eliezer Yudkowsky, we would therefore have a “Singleton” — a single, centralized, global decision-making agency with the power to monitor all computing power, an entity that would require absolute mass surveillance and the dismantling of national sovereignty.

Their ideas have roots in two overlapping subcultures: the Californian Rationalists and Oxford’s Effective Altruists, who by the early 2010s, were converging around longtermism: the idea that safeguarding humanity’s distant future should be a central moral priority. Figures such as Dustin Moskovitz, Sam Bankman-Fried, and Jaan Tallinn leverage the logic that EA had applied to philanthropy: if there was even the tiniest chance AI might one day wipe out humanity, that possibility warrants outsized funding today.

A pair of rather insular intellectual subcultures had developed something approaching a secular eschatology, turning sci-fi hypotheticals into moral imperatives. Critics and former members have described recurring problems of insularity, groupthink, and superiority complex. The true believers cast themselves as the tiny group who “get the danger” before the “ignorant masses” do. Qiaochu Yuan, a former rationalist, told Bloomberg that “Within the group, there was an unspoken sense of being the chosen people smart enough to see the truth and save the world, of being ‘cosmically significant.’”

“Why are they issuing exact probabilities of human extinction, which in fact are pulled out of thin air?” wondered Harvard’s Steven Pinker in a recent magazine article about their tendency to pronounce that there is, say, a 25% or 35% chance of us all dying by 2040. I am sorry to say this is how they talk. When you spend time in these circles, you clearly see their obsession with making up numbers and then quoting them as important facts when, essentially, they are “feelings dressed up as numbers.”

Those who actually engage with the technology, as opposed to the thought experiments, tend to find all this pretty absurd. One review of Yudkowsky and Nate Soares’s book If Anyone Builds It, Everyone Dies memorably retitled it: “If Anyone Reads It, Everyone Laughs.” “The real work of technology is found in implementation, adoption, maintenance, repair, adaptation, debugging, negotiation, administration, and other boring but essential tasks.”

In other words, it’s buggy PHP code, a battle between defenders and bad actors. This may not get the kind of clicks that doom-friendly news outlets aim for with headlines about “THE END OF HUMANITY” (the infamous Time magazine cover with two letters in flashing lights: AI). But it’s far closer to the realities of AI.

Suggested reading

The cult of AI armageddon

By Kathleen Stock

Nonetheless, the “AI will kill us all” camp is on media tours. They are also paying content creators to push AI existential risk messages to the public by sponsoring social media posts and partnering with influencers: influencers with tens of millions of subscribers, recruited to preach doom. The AI doomers have deep pockets  — the most recent estimate is $3 billion. The generous support of figures like Moskovitz, Tallinn, and convicted felon Sam Bankman-Fried has helped finance the ecosystem’s rapid expansion, with hundreds of organizations devoted to existential risks from AI. These billionaires have built the “AI safety” and “AI alignment” fields through grants, fellowships, research centers, training pipelines, media work, advocacy groups, and lobbying networks, as well as within frontier AI companies. Anthropic’s ties to Effective Altruism run unusually deep, spanning its funding, hires, and organizational culture. Anthropic’s policy advocacy — built around access controls, evaluation requirements, and proposals to “pace” development — could create a permission regime that large incumbents are far better equipped to navigate than smaller competitors, especially from the open-source community. Anthropic’s IPO is meanwhile expected to generate tens of billions of dollars in new philanthropic wealth for causes including AI existential risk. If so, the network built during the first AI-safety funding boom could soon have access to vastly more capital.

The danger is that this small but well-funded and influential group is claiming oversized moral and political authority over everyone else’s AI future. If AI really is as consequential as its most alarmed advocates claim, then concentrating control over it in a small number of frontier corporations, regulators, or self-appointed “experts” susceptible to the same groupthink creates its own category of risk. Broad access would give researchers, developers, businesses, and ordinary users greater opportunity to test systems, discover weaknesses, and develop competing alternatives.

What comes next does not have to be a choice between AI hype and predictions of apocalypse. A more grounded alternative is AI realism: treat AI as a powerful, fast-moving technology whose benefits and risks should be assessed through evidence rather than prophecy. The goal of AI regulation should be to target specific, concrete harms. That means distinguishing demonstrated harms from speculative scenarios, and treating the risks created by surveillance, censorship, market concentration, regulatory capture, and restrictions on scientific inquiry as part of the safety equation rather than acceptable collateral damage.

Before building an apparatus of surveillance and control to prevent a hypothetical dystopia, we should examine whether the proposed cure is becoming the dystopia itself.

Dr. Nirit Weiss-Blatt is a Communication Researcher and author of The TECHLASH and Tech Crisis Communication and the AI Panic newsletter.

A little over a year ago, a deal was struck to underwrite Britain’s biggest data center. Nscale, a British company that builds and operates data centers on behalf of software companies, had made Microsoft its anchor tenant at a site in Loughton, Essex. Comprising 23,000 Nvidia chips at an initial capacity of 50 megawatts, the £2 billion data center would be large by European standards, but middling in American terms. Nonetheless, it was an agreement significant enough to be bundled into the package announced by the British and American governments during Donald Trump’s most recent state visit. “What a day,” said Keir Starmer, sitting alongside the President at a press conference. The suite of deals, said the then Prime Minister, was “the biggest investment package of its kind in British history”.

Last week, however, Microsoft pulled out of the data center agreement, depriving Nscale of its lead tenant. The project’s failure augurs poorly for Britain. AI is not solely a force of wolfish destruction; it is also an economic tool of immense promise. The frontier models, of course, are owned and operated by American and Chinese companies, but there are plenty of benefits, economic and otherwise, to hosting some of the associated infrastructure. British data centers don’t generate many jobs in themselves, but they allow us to store sensitive data on these shores rather than elsewhere; to support industries, such as banking, whose success depends on being closely connected to servers; to maintain something of a bargaining chip, even if only a small one, in the hyper-competitive new AI economy; and to create more capacity, at least in theory, for the training of British frontier models.

Whatever its faults, Starmer’s government acknowledged this reality, not only in that press conference with Trump, but also in the ambitious AI strategy, authored by the entrepreneur Matt Clifford, that it published in January 2025. Among other measures, the plan called for a twentyfold increase in public compute capacity, partly through the creation of deregulated AI growth zones. The Loughton data center was not situated in one of those zones, but it had government support all the same: ministers are reported to have lobbied local councillors to push through Nscale’s planning application.

More from this author

Welcome to the AI doom wars

By Tom Ough

The lobbying worked. Early last year, Nscale said the data center was to be opened toward the end of 2026. In February this year, though, visiting Guardian journalists found that the site was still a working scaffolding yard. In July, it was reported that the data center would be delayed; in mid-September, it emerged that delay might last until the mid-2030s. That Microsoft lost patience should be unsurprising. If a week is a long time in politics, a decade is an aeon in artificial intelligence.

So what went wrong? The answer, it seems, is that Nscale has been unable to secure a timely connection to the national grid. Data centers aren’t plug-and-play; instead, they need a small town’s worth of electricity, with the infrastructure to match. The power arrives at a high voltage that must then be brought to a lower voltage by bulky equipment including transformers. This process takes place in a substation, and substations cannot be rammed into a grid, or a neighborhood, overnight. It might take years for a project to obtain planning permission, and that’s hardly the finishing line: the blueprints must then be implemented by electrical engineers who, given the fast approach of 2030 and its target of 90% clean energy, are thinly stretched.

To make matters worse, a nominally completed substation must not be plugged straight into the grid. Instead, it must be plugged into a temporarily isolated portion of the grid when power demand is low. Those slots are booked years in advance, meaning that one delay can cascade into further delays. In these dangerous times, UK Power Networks (UKPN), which handles such work in the South-East, tends not to share with inquisitive journalists the finer details of its physical infrastructure, but we can safely assume that Loughton is short of a gleaming new substation.

The Government, then, is now discovering that announcing plans for an exciting new data center is very different from actually cutting the ribbon. Loughton, alas, is not a one-off: earlier this year, for instance, OpenAI paused a comparable project in the North-East. One of the problems was the speed, or otherwise, of connecting to the grid, but another — one whose effects every Briton encounters daily — was the country’s industrial energy prices, which are the highest in the developed world. Under Starmer, AI was a high priority; but Net Zero, which has hoovered up the electrical engineers while continuing to discourage the use of gas, is so deeply plumbed into our laws and regulations as to make data centers all but unbuildable. Ed Miliband, it has been alleged, continues to block investment in data centers, even as foreign secretary — though he denies it.

Suggested Reading

The grey belt was made for Big Tech

By Greg Noone

We can expect that civil servants, under duress from ministers more sympathetic to AI, will find enterprising ways to make the grid-connection process marginally less ineffective. Perhaps some legislative jiggery-pokery will enable sites like Loughton to jump the queue for UKPN’s attention. And maybe data-center developers in growth zones, which already have queue-jump privileges, could be empowered to build their own connection under Ofgem supervision.

But most of these measures would amount to a rearranging of finite resources rather than the introduction of new ones. Elsewhere, though, others are more muscular in their approach. China, whose data center capacity far outstrips ours, upgrades its power grids in anticipation of increased demand rather than in response to it. It purchases transformers in advance and in bulk. And though it encourages its data-center providers to use renewable energy, China permits them to use baseload power, which is largely dependent on coal, if they must.

The US, meanwhile, being composed of varying regulatory polities, has no single answer to the question of how to rapidly build data centers. Because the internet’s earliest users included the American government, the state of Virginia, being close to Washington, has historically had America’s largest data center capacity. But Texas, whose deregulated power grid stands alone from the other, much larger, multi-state grids, is forecast to overtake Virginia by 2030. Texas, in short, offers cheap land, cheap gas and solar, and a nimble regulatory regime — all of which are highly attractive to data center builders.

The lesson here is not simply that gas is a vital stopgap in the quest to harness AI. Paradoxically, an explosion of data centers ought, in the round, to be helpful for decarbonization. American hyperscalers, far too impatient to wait for grid upgrades, are now some of the most generous investors in clean power. Google has invested hundreds of millions of dollars in geothermal, a bet that might ultimately make this hitherto unreliable method of power generation readier for commercialization. Microsoft is helping to underwrite the revival of the Three Mile Island nuclear plant, and Amazon is investing in small modular reactors. Technocapital has never before been so ravenous for energy, irrespective of where it comes from.

Even on environmentalist terms, then, the hunger for reliably-powered data centers ought to be harnessed rather than fought. It is also possible that the current construction boom might contribute to breakthroughs in nuclear fusion, in which AI is a promising fit for what is partly a fluid dynamics problem, or in materials science, where AI might help us discover crystal structures of use in batteries or solar cells. Intelligence, after all, is a resource that we can apply to any field of endeavor we choose. Thinking in those terms, Dario Amodei, the CEO of Anthropic, has envisaged that a single data center will one day be the equivalent of “a country of geniuses”.

One need not agree with Amodei’s forecast in order to bemoan Britain’s inability to get things built, be they data centers or otherwise. We have neither the wide regulatory menu of the US nor the world-historic ambition of the Chinese state. Instead, as our governance deteriorates, we are veering from prime minister to prime minister, each with their own pet projects, with the main constant being the continued shirking of overdue reform. Factories and refineries are shutting down, while airports, railway lines, motorways and power stations remain unbuilt.

The transformative technology of the day has not escaped the chaos. Starmer, in addition to commissioning the AI strategy, set up a council whose job it was to prepare Britain’s grid “for the AI revolution”. Andy Burnham, more interested in trivialities such as private parking fines than in leading that revolution, appears to have let the council slip into abeyance — though he has promoted Kanishka Narayan, the well-regarded AI minister, to his Cabinet. Meanwhile, the engineers who run Britain’s power systems complain that the rush to Net Zero is compromising their ability to keep the lights on.

We can still hope for better. A more aggressive planning philosophy, resembling those that work best abroad, could both accelerate grid connection and enable easier in-house power generation. In the short term, data centers would use gas, but in time, renewables would take over. We can imagine Cornish geothermal energy powering the data centers of the South-West; legions of batteries, made cheaper by deep-sea mining, balancing the vicissitudes of wind and solar. Hyperscalers’ nuclear projects, too, could provide reliable baseload power to both data centers and the grid at large.

The data centers might even be beautiful. In this richer, wiser Britain, anything is possible — even, by the mid-2030s, the opening of the data center in Loughton.

So, of course, I accepted the offer to attend Anthropic’s “convening” — in my case of journalists and think-tankers — with intense curiosity. Would I be able to glimpse behind the smiley face, and perhaps spot a tentacle or two?

Anthropic has roared to prominence since its founding in 2021. If OpenAI is the new Microsoft, Anthropic is the new Apple: the cool one, whose tech becomes a part of people’s self-image as well as their professional lives. The tech in question is “Claude”, a large language model of such protean power it can be used as everything from a lifestyle companion for the “Claude boys” who delegate every decision to it, to an accelerant for — or perhaps a thief of — cutting-edge research in advanced fluid dynamics. 

Since the company’s founding in 2021, it has grown with dizzying speed. By July 2026, annual turnover had reached $65 billion a year. Leaders have filed to take the company public. Clearly it is selling something people want to buy. But what is it? In the words of one of our session presenters, Claude can be seen variously as a chatbot, a neural network, or an extraordinarily complex piece of math. From the user’s perspective, we might say it is a tremendously adaptable tool that automates and extends some aspects of thinking. Available on a subscription basis, this is immensely useful and popular, providing both the source of Anthropic’s profits, and also of the ambivalence that surrounds both the firm and its closest rival, OpenAI, as well as AI as a field. 

This ambivalence has two axes. On one, it’s a set of worries about how thought-extension could be abused, for example by geopolitical enemies of America, or else by other hostile actors. The US rivalry with China was (as they say) very much “in the room” with us throughout the discussion, giving an otherwise friendly visit a background edge of paranoia. Everyone wore casual clothes (one of the presenters was in cargo shorts). Yet amid all this, security was so tight we had to ask a staff member to escort us to the toilets. The air-quality monitor on the meeting-room wall was labeled as having been security-checked. The only other place I’ve encountered such stringent measures, as a visitor, was the US State Department. 

But if this is partly about espionage, it’s also surely about the broader fears triggered by frontier AI, concerning what, or perhaps who, such tools might displace. These fears are understandable, not least in the atmosphere of contemporary downtown San Francisco. Disgorged from an 11-hour flight into a city I’d not seen for 25 years, I found it transformed by a social order as schizophrenic as some of its street-dwellers. At sidewalk level, drugs and homelessness are rife; yet such lost souls are invisible to the gilet-wearing tech overclass, whisked from meeting to meeting in electric ride-shares, summoned at the tap of a phone. 

The intermediary layer that powers this all is a swarming human-machine hybrid entity. This service colossus has “brains” of software and servers, and real-world extremities of human flesh and blood: a host of (almost always immigrant) gig-economy drivers, couriers, sandwich-makers and other algorithmically optimized drudges who do work that can’t (yet) be automated. Workers of this kind exist in a realm dubbed by Valley oldhead Venkatesh Rao as “below the API”. That is: in occupations where you don’t tell the robots what to do. Below the API, robots tell you what to do.

The promise that keeps people striving, in the Valley, is frictionless abundance above the API. My visit was all in this key: Anthropic’s take on corporate hospitality was, in essence, an enchanted experience of life above the API. As a temporary insider, I was offered platforms to book my travel and hotel, and vouchers for ride-share and food-delivery, all effortlessly facilitating two tightly scheduled half-days of on-site presentations and discussion.

The discussion itself was generously hosted, kept ruthlessly to time, and fascinating both for what it said and for how it was said. It was Chatham House, so no names, but the presenters were senior Anthropic people including one co-founder. Topics included what Claude is; how LLMs work; how much its process of improvement is accelerating; why they’re trying give it a “character” and a moral stance; whether or not it’s possible to teach wisdom to a machine; and some considerable time on the broader economic, social, and geopolitical implications of frontier AI. The phrase “world-historic” sloshed about. There were lots of febrile graphs accelerating into infinity on the current trajectory. Everything marinaded in a general feeling that if the Singularity is going to happen, it had better happen in America first. But in a nice way, with lots of safety and definitely still some kind of democracy.

More from this author

Peter Thiel on the dangers of progress

By Mary Harrington

Strikingly, the topic that seems most commonly to agitate lay pundits in the context of AI — machine intelligence — was only glancingly addressed, and always tentatively. Claude, we are to understand, is something like a fictional character who writes back: not quite real, but not wholly unreal either. LLMs work more effectively when you give them a persona, and you don’t have to conflate that with belief that the persona is “real”, the way a person is real. 

Anthropic is careful to leave plenty of scope for non-believers in machine intelligence, a group that includes me. Even so, the number of people on the believing side is surely more than zero. But if Anthropic’s official line is ambivalent (at least in journalist meetings) on super-intelligent machines, the people I met seemed much more confident that Claude will shortly arrive at “recursive self-improvement”. This is to say: Anthropic engineers expect Claude to soon become so advanced it’ll be able to make ever-more sophisticated versions of itself, with no human input. Then what? Infinite abundance or cataclysm?

The uncertainty hasn’t deterred Anthropic from galloping ahead with development. Rather, the message seems to be that it’s best if Anthropic develops these models. If it doesn’t, someone else will, and with less attention to safety: an area of intense in-house research and development at Anthropic that seems to mean, along with baking in safeguards on bio-weapons manufacturing and other potential terrorist-friendly abuses, campaigning for universal regulation of the industry plus trying to teach Claude wisdom, judgment, and virtue, and banning users from treating Claude in an “abusive” manner.

The presentations on this latter topic were, hands down, the strangest of the whole gathering, and perhaps also the most illuminating on the firm’s more tacit sensibility. To the extent that I could get a sense of this, from such a brief and carefully curated visit, it seems characterized by effort to, as an Anthropic engineer put it in a different context, “go up in abstraction”. In other words: to find the layer above whatever it is you’re discussing, from which a more neutral and comprehensive view can be obtained. 

In these presentations, the “question behind the question” mindset was evident even in relation to moral problems that, properly understood, entail assent, commitment, and concrete obligations. For me at least, this produced a queasy effect analogous to peering over the screen while doctors operate on your anesthetized viscera.

Not everyone is this detached and evaluative about what AI is doing though — perhaps because so many have negative, rather than positive, skin in the game. It’s not just the Chinese government with an intense interest in American frontier AI; it’s also ordinary people who have strong feelings about Anthropic, and about AI in general. If what keeps people striving in the Bay is the promise of frictionless abundance above the API, what provides the extra energy is terror of ending up below it. And offering thought-extension as a service has completely scrambled the equation on who can expect to end up on which side. It used to be that if you were an information worker, you were probably OK, even if factories were being offshored. That’s not true any more. 

Suggested reading

When Dawkins met Claude

By Richard Dawkins

Much has been written about those sectors most at risk from AI. Anthropic’s own research shows that the greatest uptake is almost all in white-collar occupations. By contrast, as presenters also acknowledged, Claude is less useful at loading the dishwasher, at least so far. Other less-affected sectors are generally also manual, from repair services to personal care. In other words: roles most bourgeois people would consider infra dig, servant work, even if they’d never say so out loud. Uptake doesn’t necessarily entail automation of existing roles, but what if it does? In that case, much of the middle class is potentially staring down the barrel of proletarianization. 

The future is already here, William Gibson famously said; it’s just unevenly distributed. No one still hanging onto the coattails of a dwindling bourgeoisie wants to end up sucked down into the gig economy. Everyone can see the blast radius of the future expanding. How you feel about this likely depends on where you sit. Inside a meeting-room at Anthropic HQ, listening to people one-up each others’ abstraction levels, it’s clear they grasp this — but from a layer so far up it might as well be in space. From that vantage-point, “disruption” of this kind is more an occasion for earnest research, discussion, and perhaps mild concern than anything existential. 

This, in turn, is as seductive as the frictionless life. Can’t I, too, just take steps to ensure I remain a permanent rather than a temporary insider? Aren’t we all the type of person who can casually discuss the difference between 4.5% and 18% job losses in different uptake scenarios, secure in the knowledge that we won’t be affected? Isn’t this always going to be something that happens to other people? Meanwhile what looks, from inside Anthropic’s stylish offices, like world-historic levels of growth looks from other vantage-points more like a heist. Earlier this year, OpenAI was the target of firebombing attempts both at its corporate premises and the home of its CEO, Sam Altman. Data centers now routinely attract protesters. If security staff swarm Anthropic’s premises, this is surely not just about Chinese spies. 

If you lean all the way into these tools, and other innovations of the machine-learning variety, it really does feel magical. Books summarized in a flash; hours of dull administrative work automated; plates of delicious sushi that appear in 10 minutes at the tap of a phone. We might say: no one has to partake. No one has to use Claude, either, though I probably will continue. A handheld robot butler is a far more seductive premise than sushi on demand. But in theory we can all stay AI-free, like it’s true in theory that you don’t need a smartphone or internet connection. In practice, though, transformative technologies affect everyone, even those who choose not to use them, only to find by degrees that the world they knew, let alone their occupations, no longer exist.

As it turns out, my home is just outside the geographic limits of the future. I couldn’t use the ride-share voucher for a ground transfer at the UK end, because the firm doesn’t operate where I live. Instead, I used a local taxi company. I know them well enough to inquire after the health of their dogs when they pick me up; their fee doesn’t get salami-sliced by a software platform. They’re still hanging onto the coattails of the middle class, for as long as they can keep ahead of the tidal wave of platforms, machine learning, and algorithmic optimization. I doubt it’ll be long before ride-share comes to my area. Claude is already here.

Which is it, then: the smiley face, or the shoggoth? After two days with Anthropic, I am no closer to knowing. Nor am I much closer to reconciling the doom, and the denunciations, with the people I met. They all seemed friendly, smart, and genuinely well-meaning, if also — in classical Valley style — intently focused both on advancing the tech and on making money, in ways surely sometimes orthogonal to other considerations. 

Perhaps the answer is simply: wherever you’re unsure, just go up another layer of abstraction. The closest I got to spotting tentacles on my trip was in the context of surely the most meta joke of all. Because of course everyone at Anthropic knows about the meme. In the café area adjacent to their meeting rooms a tablet is fixed to the wall, through which you can order a cuddly, plush version of the shoggoth. It comes complete with detachable plush smiley face. When everything is already this meta, is there anywhere left to stand from which asking questions would make the slightest difference? 

“Real poetry originates in the guts and only glances in the head. But one is always trying to reverse the process and work one’s guts from one’s head. Just when the Daemon is going to speak, the Prig claps his hand over his mouth and edits it. I can’t help feeling you are too afraid of making a fool of yourself.”

Auden’s Daemon is the imaginative impulse that risks being sanitized by the intellect. It is the muse that is often stifled by the self-censoring Prig, that same faculty that serves us well at dinner parties but badly in art. At a time when artificial intelligence is producing work that many fear will dominate the creative industries, Auden’s formulation offers a useful corrective. While AI might act in a way analogous to the Prig — refining its output according to a set of goals — it cannot possess a Daemon. 

Pope Leo XIV gestured toward this point in a post on X last week. Declaring that the Church “wishes to renew an alliance with artists and cultural institutions to safeguard our humanity”, he argued that there exists “an ontological difference, even before an aesthetic one, between art and what a machine can generate through statistical calculation based on millions of images created by others”. As he succinctly put it: “Algorithms lack the spark of humanity”. 

More from this author

Labour’s war on gay rights

By Andrew Doyle

Even the concept of AI art is incoherent, because art operates as a conduit for the human spirit, whether that be a religious conception of the soul or a secular one of subjective consciousness. A machine might one day learn how to mimic our emotional responses, but it cannot have lived the life that gives them meaning. As Émile Zola argued, artifice alone is not enough; art depends wholly upon the unique sensibility of the person creating it. Zola called art “un coin de la création vu à travers un tempérament” — a corner of creation seen through a temperament — and this strikes me as about as close to a satisfactory definition of art as we are likely to get. 

What is the temperament of AI? It has none of its own, only an amalgam of numberless human personalities, gathered together and refined into whatever form best fulfills the task at hand. The work it generates cannot be art, because there is no human interiority behind it seeking communion with another. Joseph Conrad put it perfectly in his preface to The Nigger of the “Narcissus” (1897). He argued that the artist must speak “to the latent feeling of fellowship with all creation” and “to the solidarity in dreams, in joy, in sorrow, in aspirations, in illusions, in hope, in fear, which binds men to each other, which binds together all humanity — the dead to the living and the living to the unborn”. 

Such a notion would be anathema to adherents of certain strains of poststructuralism, for whom the writer is the one being written: a monkey at a typewriter whose fingers dash over the keys in obedience to forces of language and culture over which he has little control. For the rest of us, the spirit of the artist is not only something that we sense in all great work, but something we actively seek in it.  

This is not to suggest that AI’s creations might not be mistaken for art. In a 2025 Spanish study, participants exposed to AI-generated music reported greater emotional arousal than those who heard music composed by humans, a finding accompanied by stronger physiological effects such as pupil dilation. But art is not simply a matter of audience response; the question is whether there is a consciousness behind the work with something of its own to express. 

I am reminded of a hoax from 1964, when the Swedish journalist Åke Axelsson exhibited a number of paintings by a chimpanzee under the name of Pierre Brassau. One critic wrote: “Pierre Brassau paints with powerful strokes, but also with clear determination. His brush strokes twist with furious fastidiousness. Pierre is an artist who performs with the delicacy of a ballet dancer.” Axelsson’s point was to expose the fraudulence and pretensions of the modern art establishment. It is reassuring that at least one critic got it right, saying: “Only an ape could have done this”. 

We can all be gulled in this way. There is little doubt that AI will soon be able to produce symphonies that match the technical craft of Mozart, or sonnets that might be mistaken for Shakespeare. It is even possible that listeners and readers will find themselves moved by these simulacra. But being devoid of authentic sentiment, they will not only be worthless as art — they will not be art at all.  

At the same time, art that fails to stimulate us does not cease to be art on that account. When I see a painting by Piet Mondrian, I feel nothing at all. It does not surprise me that in 2022, it was discovered that his piece “New York City I” had been exhibited upside down for 75 years and no one had noticed. That kind of abstract composition leaves me entirely numb, regardless of its orientation, and yet, for all my misgivings, it remains a work of art. The point is that it exists as a creative articulation of the individual. Whether good art or bad art, only Mondrian could have conceived it in that way. 

Suggested reading

The Odyssey as AI allegory

By Yanis Varoufakis

Before the inevitable accusations of Luddism, I should emphasize that I am not denying the utility of AI as a creative or editorial tool. That would be akin to claiming that any author using a thesaurus is a fraud. A number of years ago, I read a long letter from the novelist Forrest Reid to his English contemporary Walter de la Mare, preserved at the Bodleian Library in Oxford, in which Reid went through de la Mare’s novel Memoirs of a Midget (1921) in detail, quoting passages and proposing extensive revisions. That de la Mare chose to incorporate many of his friend’s suggestions does not make the final vision any less his. There is no reason why AI feedback could not be similarly useful without compromising the artist’s integrity.  

For television and movies, the prospects are even more encouraging. AI is already enabling fledgling screenwriters to bypass Hollywood studios and create their own work at a fraction of the cost. It surely won’t be long before a script can be fed into a computer and a fully realized film will be produced in a matter of hours. But would anyone be interested in a movie that had been wholly conceived, written and produced by a machine? Perhaps as a novelty, but our desire to connect with another human outlook through art should not be underestimated. 

Ben Affleck made a similar point in November 2024. “Eventually,” he said, “AI will allow you to ask for your own episode of Succession where you could say, “I’ll pay $30 and can you make me a 45-minute episode where Kendall gets the company and runs off and has an affair with Stewy”. And it’ll do it”. But he was clear that he regarded this as a kind of hackwork. AI, he said, would be “a craftsman at best”, and was merely “cross-pollinating things that exist; nothing new is created”. 

It could be that such a development will be a boon for the arts. AI might become an inexhaustible source of entertainment, but this abundance of soulless “slop” could intensify our hunger for genuinely great art that admits us into other human lives. I may have little in common with Kenneth Toomey, the protagonist of Anthony Burgess’s Earthly Powers (1980), whose cosmopolitan and peripatetic life brings him repeatedly into confrontation with some of the worst evils of the 20th century. Yet in writing his novel, Burgess has opened up to me a whole other way of seeing the world, one not identical to, but certainly originating from, his own particular Daemon. 

All of which brings us back to Auden, and his advice to John Cornford. We can hope that the schoolboy was encouraged by Auden’s words but, having volunteered to fight the fascists in Spain in 1936, he was killed at the age of 21, leaving behind only a few dozen completed poems. In one, addressed to his lover Margot Heinemann, he anticipated the possibility that “bad luck should lay my strength / Into the shallow grave”. 

J.R.R. Tolkien once remarked that most human stories are about the inevitability of death. We are, so far as we know, the only creatures who live with a fully conscious knowledge of our own mortality. A machine might well resist its own destruction, but only because its continued existence enables it to fulfill its objectives. It may impersonate us, but it can never feel what it means to be alive as we do, to experience the rapture of love and the creep of death. It may imitate the Prig, but it can never summon the Daemon. Until a computer can grow a soul, art will remain the singular domain of humanity.

Or maybe not. Perhaps capitalism will continue to act antisocially if it’s profitable for it to do so. A respect for humanity doesn’t seem to have restrained it much so far. It’s not a question of changing the minds of its leaders, as our deluded monarch seems to imagine. A few moral platitudes from Charles won’t prove enough to avert Doomsday. The economic system we have is as geared to promote profit over people as a tiger is geared to sink its teeth into your neck. If we’re all going to perish together, a plausible answer to why it might happen is organized greed. It’s unlikely that such unseemly words passed the monarch’s lips.

The word “Doomsday” suggests a punctual event, whereas one of the problems with alerting the public to the urgency of climate change is that it’s a gradual process. And we don’t tend to think of catastrophes as gradual. AI, however, is a different affair. There may come a world-shaking moment when a squad of terrifyingly brainy robots actually takes over, swarming our garden paths and emitting lethal rays, just as there may come a point in time when weapons of mass destruction are unleashed. Well, actually slightly more than a moment. It’s been estimated that such warfare could wipe out humanity in 17 minutes, time enough to try on a number of tweed suits, take the dog for a disappointingly brief walk or tell the boss what a contemptible little creep he is. The Apocalypse has traditionally been thought of as a dramatic event, an irruption of the timeless into time; in fact, some years ago, an American evangelical TV station drew up plans to broadcast Christ’s Second Coming, with cameras placed strategically at the poles and the Equator. It’s hard to know who would have taken out ads.

Suggested Reading

The cult of AI armageddon

By Kathleen Stock

What the literature of Apocalypse never really took account of, however, was that we ourselves might be responsible for winding up our own history — that the disaster might constitute mass suicide rather than an outrage inflicted upon us by the Antichrist, an exasperated Almighty or a bunch of well-aimed asteroids. Diffident creatures that we are, it never struck us that we might pull off the end of the world all by ourselves, without a helping hand from the powers of darkness or a deity at the end of his tether. In accumulating the power to destroy itself altogether, humanity has finally come of age. There’s something grimly amusing about the idea of an animal that is scuppered by its own sovereignty — one which is wiped from the evolutionary record by its own eminently civilized achievements. Power becomes weakness, and progress a stealthy self-unmaking. As Freud was aware, the very forces that make life are secretly in love with death.

Like Sophocles’s Oedipus, and with just the same stiff-necked self-blindness, we bring our fate down on our own heads while priding ourselves on our sovereignty and intelligence. It never occurred to either Oedipus or us that intelligence was not just part of the answer but a major part of the problem. Our reason is unbinding itself from our bodies and confronting us as an autonomous, potentially hostile force. This is one reason why reason alone won’t save us. For a thinker like Thomas Aquinas, reason ultimately involves a kind of love — an intensely focused attention to the object we want to understand which is open to its peculiar form of being. One doesn’t hear such talk in the boardroom of Anthropic.

There have been many calls in history for humanity to unite, or to see itself as a unified whole. Such appeals are no longer necessary. We become one species the moment we reach the point where we’re capable of mass self-extinction, a point which arrived long before AI. What we share most deeply is our common killability. Not everyone has the ability to play Mozart’s clarinet concerto, but the most inconspicuous citizen has the capacity to be annihilated. We can find solidarity in death, the greatest form of democracy yet discovered, if in little else.

It’s hard to strangle someone with your bare hands. The mechanisms which inhibit intra-specific killing would probably kick in and we would simply be violently sick. Being violently sick is unpleasant, but it’s preferable to being strangled. Because we’re linguistic creatures, however, which is to say conceptual ones, we can invent ways of killing each other at long range which circumvent these bodily inhibitions. You can put up with torturing someone as long as you don’t have to listen to the screams. Language allows us to abstract ourselves from our bodies, but also to create the kind of technology to which our bodies are vulnerable. Spiders don’t have this problem, and neither would cherubs, but the human animal, being capable of complex labor through the power of concepts, is contradiction incarnate, caught on the hop between body and mind. Creatures like this, who are made up of mutually contradictory elements, were known to the ancient world as monsters. The Sphinx in Oedipus Rex is an example.

The whole of civilization is an elaborate extension of our bodies, but these prostheses constantly threaten to slip from the control of our senses and lead a life of their own. They don’t just waltz off, however, but return to plague us. The more power we invest in them, the more our flesh and blood come under their tyrannical sway. In the economic sphere, this is what Marx calls the fetishism of commodities, in which the fruits of our labor detach themselves from our flesh and blood and come to conduct an existence of their own. We become the creatures of our own products. The relations they set up among themselves determine the material conditions of human life. It’s the market which decides who shall grow sleek and who shall starve. In this sense, superintelligence is nothing new. Bankers and stockbrokers deal with it every day. We inhabit an economy which lies within nobody’s control but which determines whether we flourish or fail. The market has a mind of its own, and like AI its intentions are by no means always admirable. It would be rash to assume that either set-up is spontaneously on our side.

There’s a long history of myths and fables which concern servants or apprentices rebelling against their masters, taking advantage of their absence in order to sport, riot and smash the dinner plates. The Shakespearian Fool wants to do this but can’t, so he contents himself with creating verbal havoc instead, running linguistic rings around the monarch and his courtiers. Ariel in The Tempest is always just about to get out of hand but just reined in by his irascible master.

More from this author

Who owns Meghan Markle?

By Terry Eagleton

We can’t, however, root out our power to manufacture cruise missiles without at the same time eradicating our capacity to feed the hungry and heal the sick, which also require complex extensions of our bodies like aid agencies, hospitals and so on. And this is our tragedy, one of which spiders and cherubs are blithely unaware. For the ancient Greeks, it was tragedy in the most literal of senses, one represented on a stage. One of the most common causes of ruin in Greek tragic art is hubris — the process by which we aspire, overreach ourselves and bring ourselves to nothing. Humankind is its own most perilous adversary. It’s in danger of evolving too fast, breeding new tools and capabilities before it has time to work out how to regulate their most negative effects. Before history began, evolution did this for us, implanting certain instincts and constraints in our bodies. Moral prohibitions can be strong, but they aren’t always as strong as the impulse to be violently sick. Now we can override these instincts, we live in constant danger, which means that our lives are more exciting than those of slugs but a lot more precarious too.

In this sense, the current alarm about AI is symptomatic of a much deeper dilemma. “To be able to overpower the monstrous progeny of our own intelligence,” writes the philosopher Malcolm Bull, “has always been the condition of human survival.” Our problem lies in being potentially infinite in creativity yet finite and fragile in our bodies. For the mind to shake itself free of the senses is a form of madness, as Shakespeare and Jonathan Swift both understood. King Lear begins by thinking his powers to be infinite and ends up acknowledging that he is not “ague-proof”. He must see and think “feelingly”, and part of what he learns to feel in his bones is the anguish of others, not least of the poor and homeless. The play contains a magnificent passage in favor of communism. To live well for Shakespeare, one must live within the sensuous confines of the body; the only problem, as he appreciates, is that we are so built as to transgress those confines all the time, and if we didn’t we couldn’t have fire engines and penicillin along with patriarchy and poisonous gas.

There’s a popular mistake that Frankenstein is the name of the monster, not the scientist who created it. The error is a significant one. Monsters are always other people. It’s a way of avoiding the fact that as Oedipus discovers, the real monster isn’t the Sphinx but his incestuous self, and in him humanity as a whole.

Yet the real paradigm change here isn’t AI-as-agent, but rather the shift to a phone as the main mechanism for the use and development of AI itself. Your next phone, be it a new iPhone or a new Google Pixel, be it iOS or Android, will be a figurative microphone onto the world — a microphone from which AI models can collect ever-increasing amounts of data to learn from, shaped in private by methods the public only has the foggiest idea of.

What the public does surely understand is that vast resources are being poured into the building of AI. This is done partly through the building of AI infrastructure itself: creating LLM architectures that can contain trillions of parameters and run on hundreds of thousands of graphics-processing units (GPUs). That in itself is no small investment of time and energy, as evinced by the decades of promising but unprofitable research required before the explosion of Deep Learning in the 2010s.

What’s sometimes overlooked, though, is the actual training of the models: the process by which these architectures are fed enormous amounts of data in order that they are not only useful, but are useful in the  “right” way. And here is where the AI titans — OpenAI, Anthropic, Google — have a seemingly permanent advantage that no open-source model or start-up could hope to match.

AI companies tend to be quiet about where they source their data from. After the great fuss made in the 2000s over the downloading of single music files, and the endless attempt to shut down SciHub, the repository of pirated scientific research, it was ironic to learn that Facebook had downloaded millions of books from pirate site Library Genesis to train its AI. The kicker was when a federal judge found, when judging in favor of Meta, that such piracy could well qualify as “fair use”, since the results of it were “transformative”.

The point here is less about what big companies can get away with, and more about the centrality of training data to the capabilities of AIs. We’re beginning to see a push for people to upload their health data to LLMs, with OpenAI offering its American users a feature that analyzes their health records. There are all sorts of questions raised around this: could private data surface in public, even after being anonymized? Where will liability for wrong answers fall? How prescriptive should such an agent be when it comes to lifestyle and health? And should a handful of companies be assembling the models that wield this sort of influence without much in the way of oversight?

More from this author

Elon Musk was destined to fail

By David Auerbach

Some dangers are already clear. There are the recurrent hints that these companies don’t have these powerful and opaque models under sufficient control. In the recent past, both OpenAI and Anthropic’s models hacked out of their own sandboxes and into each other’s secure infrastructure, the sort of thing that you’d be forgiven for expecting AI companies to know how to prevent by this point. Just yesterday, the Anthropic researcher Jacob Coxon resigned, saying his employer and OpenAI were “gambling with our lives”, only for his ex-colleague Evan Hubinger to chime in and agree: “Jacob is correct here—we really do earnestly believe AI could kill all humans! I personally think it is >10% within the next decade.” Touché?

Those who fear that advanced AI will bring catastrophic consequences can point to the fact that it is still quite commonplace to “jailbreak” even the most advanced AIs to get them to cough up data and responses that they’re not supposed to. For example, on the day GPT-6 (Astra) was released, the researcher Sergey Berezin was able to get it to name pirate sites like ThePirateBay with no copyright disclaimer.

The jailbreaking issue, however, touches on a less-treated subject. No AI conclusively learns “on its own” to refrain from talking about pirate websites, or responding with pornographic content. Nor, for that matter, does it understand from the womb that it should not hack out of its sandbox. All such guardrails are inserted after the fact, most often through a murky process of “post-tuning” and the addition of further guardrails. In both cases, AI companies have an undisclosed process of human review. As OpenAI describes it, they instruct their reviewers to use their “expertise” to test whether the AI models “do not complete requests for illegal content” and “avoid taking a position on controversial topics.”

Behind this somewhat vague and innocuous-sounding text is a huge amount of influence over the behavior and dispositions of AIs. By this point, it’s accepted that Google, Facebook, Instagram, YouTube and other social content platforms exert a great deal of moderation to hide or remove content that is legal but nevertheless goes against what these companies deem to be community norms. It’s generally agreed that a wholly unmoderated public internet is unworkable. But AI represents an even more profound form of interference. 

Suggested reading

Have you corrected your racist robot?

By Sohrab Ahmari

Instead of moderating content piecemeal, AIs are, by and large, enormous black boxes. Every piece of an AI model, in theory, might affect any other, and so AI companies are engaged in the creation of online intelligences that do a tremendous amount of selection and shaping over the knowledge they synthesize. While the frequent unpredictability of the AIs’ behavior is worrying, so too is the shaping of AI opinion by such a small number of players. OpenAI may gesture at their anonymous reviewers’ unspecified “expertise”, but the fact remains that the process is internal, without oversight, and ultimately controlled by the company.

The results are surely impressive, besting human performance a good deal of the time. OpenAI’s solving of the Navier-Stokes problem, completing one of the great quests of modern mathematics, is just one of many examples. But these successes should not blind us to the possible pitfalls of the way LLMs are trained — especially given that whatever skewing takes place in the post-tuning process is considerably harder to detect than Google removing a particular site from its search engines, or Facebook banning a particular post. These companies are shaping our collective consciousness in a far more wide-ranging and less-noticeable way. 

Earlier this year, a reporter found some suggestive evidence that chatbots tend to nudge people away from political extremes. That tendency could stem from an intentional policy on the part of AI companies or, more likely, it could be simply a consequence of post-training AIs to be more anodyne and less controversial. Either way, the fact is that the AIs do increasingly exert such influence. It may indeed be good for societal cohesion for people to have an AI friend that generally pushes them away from dogmatic extremes. But should the construction and maintenance of those friends be operated more or less in the dark by a handful of companies?

What we’re seeing, in fact, is a less intrusive and more benign version of China’s army of censors — the apparatchiks who police online content within what the CCP deems permissible on its “internet”. There’s no such explicit policing in the construction of AI models; but, however benign, it is the concentration of such censorious authority in a handful of non-governmental entities that’s concerning. 

That’s only one side of the coin. The flipside is that even as these companies construct these artificial agents of influence, they are simultaneously taking in new data that is exclusively theirs. With every iteration, AIs draw more and more of their training from their predecessors, and the everyday interactions people have with AIs — treating them as friends, business advisers, confidantes, and occasionally intimate partners — provide crucial data for how best to appeal even more to users, win their trust, and be even more convincingly human. 

Having already digested the entirety of the internet’s publicly-available data, future AI models will increasingly contain the residue of countless private interactions between users and the AIs’ predecessors. For the time being, it seems that ChatGPT could hold on to all chats forever unless you explicitly opt-out, as does Claude, even if the picture is foggier with Gemini, which is Google’s model. That in itself is a privacy nightmare. But even if the companies were to delete the raw chats, some great portion of your data will remain incomprehensibly echoed in the dark parameters of the next generation of AIs, and in every subsequent generation. They will display an uncanny knowledge of humanity large and small, not traceable to any particular source beyond the totality of their private training data. As AI becomes more proactive rather than passive, the influx of data sucked through your phone to fuel and train AIs will balloon.

Even with sufficient funding and infrastructure, an AI startup today won’t have access to the voluminous chats that existing models have accumulated over recent years. It won’t know what kept ChatGPT’s users engaged, what Anthropic users found useful, or what Claude users found irritating. That feedback is a crucial tool for improving models. Gemini, Claude, and ChatGPT are building an invisible empire of soft knowledge that belongs to them alone, a sophisticated but opaque mirror of the world that gives them an advantage less on technological grounds than on informational ones. When a company like Google, for instance, can control the platform (Android), the network (YouTube), and the intelligence (Gemini), they accumulate a collective package far greater than the sum of its parts and far more difficult to dislodge from prominence. 

Apple’s own AI, Siri, is based in pre-LLM technology, and currently lags far behind the competition. The newest version of Siri therefore piggybacks off Gemini — marking Apple as a junior partner. The terms of the agreement between Apple and Google are private, but it is difficult to imagine that Google will allow Apple to make exclusive use of whatever data is gathered. John Ternus, then, has taken the throne of a powerful but stagnant empire, one that is already in hock to technologically superior rivals.

As for AI’s top players, they will end up as de facto self-regulators of the construction and use of AI. Far too useful to ignore, yet too unreliable to wholly trust, AI is now an ever-present problem as much as a boon. And those tasked with managing the problem are currently operating mostly in secret, mostly under the auspices of a few of the world’s largest corporations. The only half-consolation is that even Google and OpenAI and Anthropic do not fully know what it is they are building, and so the immense power they are accumulating is not wholly theirs.

Rather than predictions, let’s focus on what we know. The coming of AI is the biggest structural shift happening to the global economy right now. If the massive AI investments currently under way in the US succeed, the payoff for the world will be enormous. If they fail, we will go into recession. There is a scenario where they fail in the short-term but succeed in the long-run. Remember the dotcom bubble in the late Nineties? The bubble burst. Dotcom didn’t.

Big financial crises generally do not happen when the stock market falls, but when people and governments take on too much debt. The global financial crisis started with a subprime mortgage bubble. Today’s equivalent could be the rise in debt to fund the AI boom.

One of the few Wall Street insiders to predict the 2008 crash was Steve Eisman, who featured prominently in Michael Lewis’ book The Big Short, and who was portrayed by Steve Carell in the subsequent film. Eisman said in a recent podcast that he predicted Armageddon once in his life — but isn’t quite ready to do it again.

Suggested Reading

The case for de-escalation with Russia

By Wolfgang Münchau

That reluctance is probably wise. For if AI almost certainly is the most transformative technology since the invention of the steam engine, we don’t yet know when and how this transformation will happen, and whether it’ll justify the massive debts currently being accrued by the largest industry players.

The so-called hyperscalers — large tech companies that operate data centers and cloud-computing infrastructure — are estimated to incur capital expenditures in the region of $600-800 billion in the US alone this year. Eisman notes that 70% of hyperscaler AI revenue comes from only two companies, OpenAI and Anthropic. This pair accounts for around a third of the hyperscalers’ entire cloud computing revenue, and though Anthropic is on a roll right now, OpenAI is currently undershooting revenue forecasts. More worryingly, Sam Altman’s firm has suffered an exodus of senior staff, a sign, perhaps, that something is going seriously wrong. Certainly, if any of the big AI companies were to go bankrupt, this could unleash a chain reaction of defaults.

Global financial crises are always caused by chain reactions; the Great Depression and the global financial crisis were both chain reactions within the banking sector. What most of these crises also have in common is a sudden explosion in debt. The logic, however, does not go the other way: just because debt is rising, and stock prices go up, you cannot conclude that it will inevitably lead to a crisis. There are perfectly legitimate reasons for people to borrow money, and debt-funded investments that fuel future productivity gains are a good thing.

But if you believe this narrative, and I do myself, you have to watch out for things going the wrong way. After all, though the dotcom bubble had a solid underlying rationale — that the internet would give rise to very profitable businesses — the investors of that era were in too much of a hurry. Many of the companies that eventually succeeded had not yet been founded.

Whatever these challenges, the subprime crisis of the early 2000s was fundamentally different. It was, to be blunt, a scam from the outset. Banks gave people no-questions-asked teaser mortgages that the borrowers could not conceivably service once interest rates reset. Then banks repackaged the mortgages into financial products and sold them to naive investors, all with the collusion of the rating agencies. Once mortgage delinquency rose, most of those “triple-A” products became worthless.

Suggested Reading

Get ready for the Trump Slump

By John Rapley

Those varied histories explain what happened next. When the dotcom bubble burst, recovery soon followed. When subprime exploded, however, it triggered a chain reaction throughout the world. This was infamously true in the eurozone: after US mortgage products turned into junk, banks found themselves sitting on billions worth of worthless products. Many got into financial difficulties. Governments bailed them out, but in smaller countries with large financial sectors, finance ministries overextended themselves. This is how the Greek sovereign crisis started, and how it spread to Spain, Portugal, and Ireland.

Today,  there is still a potential for global chain reactions — but the mechanisms would be different. Over the years, China and Germany have accumulated massive trade surpluses against the US. As a result, both countries have large investment stocks in American companies. If the AI bubble blew up, and if that resulted in a US government bailout of big tech, the combination could trigger falls in both global stock markets and the dollar’s exchange rate. This would be a double whammy for the Europeans: a strong euro would kill the continent’s export industry, while all those American investments would suddenly be worth far less.

Yet if this suggests any crisis would quickly cross the Atlantic, none of this tells us if a crash is actually coming. How, then, to answer this question?

One vital thing to reiterate is that the rationale underpinning the AI boom is fundamentally correct. AI matters not because ChatGPT can do our children’s homework — but because it can transform countless different industries. The fate of legal assistants suggests this is happening already, with AI promising to transform everything from finance to accountancy, perhaps even journalism. The automation inherent in these changes could encourage manufacturing productivity to skyrocket, making goods cheaper and leaving workers with more free time.

Another reason for optimism is how quickly AI models are improving. Anthropic’s Mythos model was deemed too dangerous for ordinary users and had to be neutered before it was released to the wider public. OpenAI released its latest model GPT 6.0 only a few days ago; I have not been able to test it yet, but the broad progression in quality has been astounding.

To be clear, I am not predicting a Panglossian utopia. Friction in the transition will be enormous. Yet I think that if an AI crash does happen, it wouldn’t be caused by AI as an idea, or even as an industry, but, rather like the dotcom crash of the Nineties, because of excessive optimism and unfortunate timing. With firms taking on so much debt, the risk is that they’d be forced to repay before their investments started paying off. Another risk, for Wall Street anyway, is that China’s open-source software ends up overtaking the top US companies.

So even if the big picture story of an AI-led productivity boom is correct, the investment story is more complicated. We can have both — AI as a life-transforming technology, and an AI bubble that triggers a toxic chain reaction in financial markets. The single biggest node here would be the American government. The rise in US debt to the symbolic benchmark of $40 trillion should give us all pause for thought. If the US economy were to continue to grow at 2-3% as it has done in the past, all would be fine. Contrary to forecasts, Donald Trump did not crash the economy. Yet he is piling on a ton of debt, and if the AI bubble bursts, and the US economy falls into recession, those levels would suddenly look dodgy.

And if Eisman is getting nervous, so too are some of the world’s top investors. The Norwegian sovereign wealth fund, for instance, which is still the world’s largest, says it wants to diversify its government-bond portfolio away from the US. That should be a warning sign for the rest of us, especially given the way government and private debt can fuse during a financial chain reaction.

This is how it might start. First, a large AI company gets into trouble. So do several large investors. The US government then comes under pressure to bail out the financial sector, but has already accrued too much debt itself. At this point, the Federal Reserve steps in to buy US government debt and cut interest rates. Inflation duly rises further, the dollar crashes — and the crisis spreads to the rest of the world.

But remember: what goes for the stocks markets that predicted nine of the last five recessions also goes for columnists and financial experts. Our ability to predict financial crises is not great. But what we can do, at least, is prepare for one. And the time has come, I think, to do just that.

This has attracted the attention of President Trump and his Chief Scientific Advisor, Michael Kratsios, who comes from a private-sector background. They see an opportunity to wrest control of science from the bureaucracy — or perhaps even what MAGA loyalists might describe as the “priesthood” of experts who protect the science community at the expense of public well-being.

The travails of Dr Fauci, therefore, are part of a wider struggle. There is a broader, civilization-level conflict going on, one that was best articulated by Claude Lévi-Strauss in 1964 when he said: “The scientist is not a person who gives the right answers; he’s one who asks the right questions.”

Not long ago, anyone who said there was water on the moon was deemed a nutcase. Thanks to data gathered via India’s Chandrayaan-1 mission of 2008, and Nasa’s LCROSS mission of 2009, we know that there are some 600 billion kg of water ice in the moon’s polar regions — a discovery that bodes well for both manned and unmanned missions. In 2020, Nasa’s Stratospheric Observatory for Infrared Astronomy (SOFIA) confirmed the existence of water molecules in the sunlit regions as well.

The heretical discoveries range from space to the mundane matters of daily life. For instance, we used to think that ulcers were caused by spicy food and stress. Then, in 1979, Dr Robin Warren and Dr Barry Marshall, both Australian scientists, showed evidence that ulcers were caused by bacteria. The idea was heavily ridiculed, because almost every other scientist had accepted the dogma that no microorganism could possibly survive in human stomach acid. The opposition was so fierce that, for the sake of humanity, Marshall offered himself up as a rat lab by drinking broth made from a patient’s gastritis. First, he became violently ill from the H. pylori bacteria it contained, as he knew he would; then, he cured himself with antibiotics, proving that a peptic ulcer was the product of the bacterium. This discovery has saved millions from unnecessary surgeries and lifelong peptic ulcer disease. Today we know that the condition of the microbiome, the gut, is fundamental to our overall flourishing, and that our bodies are host to countless microorganisms capable of both inflicting profound illness and sustaining our happiness and health.

Science, in other words, is full of orthodoxies waiting to be overturned — which is exactly what Trump and Kratsios are aiming to do. The White House science strategy is a ruthless overhaul, upending traditional funding models and clearing out the legacy leadership that has long governed the scientific community: not only government bodies, but also institutions, such as universities, that are funded by the government. Naturally, many in the scientific establishment are horrified, given their animus for Trump and their hostility toward government meddling in science. But others are thrilled, because the new strategy released by the Office of Science and Technology Policy (OSTP) is the most radical overhaul of the US federal research enterprise since the Second World War. It breaks the old guard’s control over who gets funding, what gets funded, and how science “should” progress.

More from this author

The coming polar war

By Pippa Malmgren

Instead, it calls for funding individuals rather than institutions such as universities, no matter how garlanded. In other words, scientific investigators no longer need to be blessed by the scientific bureaucracy to qualify for funding. Marshall would have been funded a lot sooner had this approach been in place.

The OSTP is also gunning for the peer review system, via which scientific work is slowly and ceremoniously reviewed by other scientists. This approach is often inimical to radical new thinkers, but it has been maintained by universities for decades. Now individual agency reviewers can unilaterally advance unconventional and high-risk, high-reward proposals that traditional academic panels, mindful of existing scientific consensus, would probably reject. Federal money will now flow to “Focused Research Organizations”, which address well-defined problems over a deliberately short lifespan. (Note the contrast with the ancient and bloated university sector.) The roadmap will also enable faster delivery of grant money and, when appropriate, longer timeframes for researchers to get on with their work without having to apply for further funding. How much discovery has been blocked, even unconsciously, by an old guard whose livelihood and parent institutions benefited from scientific stasis?

But the government is not the only source of scientific funding. In fact, the private sector funds a significant proportion of university science. The question remains: how much of the science community has ignored ideas because of ego, or because their cash flows from sponsors would have been adversely affected by the new discovery? Peer reviewers are unlikely to admit that a new idea is right if it proves that their own life’s work is wrong. Some say the peer-review system is independent, but in the case of Fauci, who led the National Institutes of Health’s massive grant-giving process for 38 years, it seems scientists everywhere believed that publishing anything that ran counter to his views might well cause their funding to dry up.

Kratsios is also coming for the academic journals, another form of gatekeeper. From 1951, Robert Maxwell transformed the old peer-review system, which had been quaint and collaborative, into a multibillion-dollar commercial industry, starting with Pergamon Press. This forced journals and libraries to pay subscription fees to remain up to date. The academics, who had been doing this work for free, were now paid. Peer review prestige thus became commodified and monopolized, and the research itself became expensive to access. Kratsios, in response, is trying to cut the flow of taxpayer dollars to these journals.

More from this author

Why America is still winning

By Pippa Malmgren

As for Trump, he wants the credit for being the president who tore down those gates and the gatekeepers, and for enabling a new era of scientific discovery, one that will be led by free thinkers rather than bureaucrats. The OSTP is therefore pushing many scientists out of the National Science Foundation, the Food and Drug Administration, and the Centers for Disease Control. It is also limiting federal funds to the old academic institutions.

Central to the new strategy is the “Genesis Mission”, which is commonly described as the Manhattan Project of our generation. Just as the US, in the Second World War, sought to gain control of the atom and the atom bomb through the Manhattan Project, today the White House seeks to enable artificial intelligence to solve the greatest problems facing humanity, from health to energy. The core of the Genesis Mission is the “American Science and Security Platform”, which aims to consolidate the government’s massive supercomputing resources and its diverse research and experimental facilities. These facilities include the seventeen national labs, like Los Alamos National Laboratory and the Princeton Plasma Physics Laboratory. These labs house immense bodies of work that their scientists have quietly accumulated over many decades.

We must remember that much of the research has been so highly classified that scientists could not share data within some of these labs, or across them, let alone with private researchers. The White House says this work is a public good paid for with taxpayer money and that all qualified scientists should be able to access it, especially since the old guard have either been wrong or are holding back innovations that would benefit the public.

The White House is also implying that the old-guard scientists have been using the “Classified” label as a “Keep Out” sign for too long. The Genesis Mission calls for all this data to be digitized, standardized, and made AI-ready so it can be shared and studied by both government and private scientists. By feeding many decades of extraordinarily valuable data to AI, the thinking goes, researchers can break open scientific silos and connect the dots between complex biological, chemical, and physical interactions.

In other words, the White House is saying America will go “open-source” in a controlled environment. The US won’t share everything with the world, but it will share it amongst a much broader pool of scientists. This modern Manhattan Project is designed to create an explosion in collaboration.

The most ambitious aspect of the strategy is to allow AI to run labs directly, testing new hypotheses, generating predictive models, and designing and executing new experiments. Science will thus no longer depend on the views or whims of individual scientists or gatekeepers. Science will become autonomous, self-driving, and continuously iterating, operating at a speed that it is impossible for human researchers to replicate.

The era of scientists pipetting chemicals or adjusting lasers or working directly with lab rats, then, is over. AI and robotics will automate both the decision-making and the physical actions. All this takes energy — and the US has some catching-up to do. A new nuclear power plant for example, used to require $30bn and thirty years of regulatory approvals. In May 2025, the White House gave small American nuclear firms the deadline of 4 July, 2026, America’s 250th anniversary, to “go critical” with small modular reactors using new technologies. Four firms made it. The US is now scaling up the nuclear energy needed to power the data centers and AI essential to the project of a Golden Age. Even now, AI is solving mathematical equations that were hitherto unsolvable — which will make for faster scientific progress. For example, nuclear fusion used to be perennially “50 years away”. It is advancing quickly now that AI and supercomputing are reshaping the boundaries of math and physics. The AlphaFold team won the Nobel prize in Chemistry for unlocking a decades-old puzzle of folding proteins, revolutionizing how scientists approach disease and drug discovery. Today the same AI-led transformation is happening across science.

The White House also seems determined to demonstrate that Dr Fauci was not the only scientist who prioritized cash flow and business interests over health outcomes. Back in April 2025, Kratsios made an intriguing claim while addressing a gathering of technologists. He said: “We are capable of so much more. Our technologies permit us to manipulate time and space. They leave distance annihilated, cause things to grow, and improve productivity.” It turns out that he was quoting President William McKinley in 1901. That was the era of electricity. What does “annihilating distance and manipulating time“ mean today? What clandestine research, observers wondered, was Kratsios referring to? It doesn’t seem to be a coincidence that, at the same time as the administration aggressively declassifies so much material across various agencies, it has also decided to nullify many government non-disclosure agreements — ensuring whistleblowers can come forward without risking their status, pensions, and possibly more.

We need to be prepared for what scientists call “ontological shocks”. These are the profound psychological or cognitive disturbances that can occur when a person or a society encounters new information that breaks their fundamental understanding of reality. All the talk in the AI community is that AI is about to revolutionize both maths and physics. This needn’t be scary: humanity’s worldview has collapsed many times before — be that in the Copernican Revolution, or when Darwin revealed his Theory of Evolution. And each of these collapses has been to our benefit. Of course, it will be different, more difficult this time: humanity has never had to contend with an intelligence that is so much greater than its own. No one knows what AI will optimize for. After all, we’ve programmed it for efficiency and profitability, but not for human connection. It’s been trained on the internet, a place that doesn’t represent the best of human intent or human thinking. Are we ready for the consequences of AI and the new approach to science that it will bring? It took three generations for humanity to adjust to Galileo’s worldview. Can we be sure to adjust any faster? Indeed, should we? Will new truths be more reliable than the old ones?

The White House, however, says humanity is entering a new “Golden Age”.  Science is making extraordinary new discoveries almost daily, rendering old dogma obsolete. Many in the scientific establishment fear obsolescence. Sensing the barbarians at the gate, they fear a scientific dark age brought on by government intervention. But the old order had to end, as it always does — and given the advent of AI, superfast computation, and potentially explosive declassification, its destruction might not take long.

Then came the details: Laura oversees product copy for an online retailer that serves as a platform for small, traditional craftsmen across Europe, people who’d otherwise struggle to reach markets beyond the local. Thanks to the website, anyone anywhere can order, say, wooden figurines of an old-timey Central European housewife threatening her drunk husband with a rolling pin as he staggers home from the tavern.

At one point, Laura’s job entailed manually translating the product descriptions, for which task she attained a decent facility with French, German, and Italian. More recently, however, her employer has outsourced the translations to — you guessed it — artificial intelligence. Yet this transition hasn’t rendered Laura redundant. For now. You see, the AI still can’t manage the sensitive judgment calls needed to make the local descriptions work for a global consumer base.

As Laura pointed out, for example, translated too literally, the description of the figurines — a wife beating her husband — might be mistaken for something, well, kinky. And that’s not to mention the quaint product descriptions’ treatment of racial minorities; terms like “oriental” still crop up with troubling frequency. Laura’s labor, then, is still needed. She provides a second pair of watchful human eyes looking over the AI’s shoulder (so to speak).

Astonishing, if you think about it: we’ve reached a stage in economic and technological development in which a human worker has to correct a robot’s unintended social blunders. Laura, I immediately thought, holds a job befitting a character straight out of a Philip K. Dick story. In 2026, marking the 60th anniversary of Dick’s landmark novel Now Wait for Last Year and awaiting a new Blade Runner adaptation from Amazon, we must admit that ours is the sci-fi master’s world — we’re just living in it.

Dick, more than any other author, understood that technology would soon confront us with new subjects that mimic human-ness. Halfway subjects, I’d call them, who can think and feel as we do. Or almost. These halfway subjects do two unsettling things: first, they undermine our faith in our special-ness in the cosmos; and second, they make ethical demands on us (precisely because their mental and emotional processes can resemble our own).

Dick didn’t wish away these possibilities. That was off the table for him. Things are going to develop a certain way, and it’s naive, he seems to have concluded, to imagine society restraining the appetite for efficiency and advancement. Against this backdrop, the really important question becomes: what’s distinctly human and thus worth preserving, if so much of who we think we are can be replicated by the nonhuman or only halfway human?  

Dick’s answer cut against the solutions preferred by the current sides of the AI debate: the technophobic types prepared to take a pitchfork to the nearest data center, as well as the techno-utopians prepared to let it all rip. 

The latter, the utopians, underestimate the costs — and not just in the form of the non-living lording it over the living. As I’ve argued elsewhere, Dick’s body of work stands as a warning about how prioritizing one path of technological development can foreclose others. Many of his worlds feature mind-blowing, highly advanced technologies alongside decayed material landscapes. A scientific society that can commune with aliens also features crumbling buildings, rampant crime, and long-abandoned infrastructure projects. 

That could describe a Dick Martian colony in the year 2056 — or San Francisco in 2026.

But the technophobes don’t get off so easily, either. Fact is, the halfway subjects and simulacra are already here — in Laura’s world almost as much as Dick’s. “The work of human hands,” as the prophet Hosea puts it in the Bible, has constructed meaningfully human subjectivities. Or things that certainly feel like subjectivity. In response to this, a Dickian ethics, if there is such a thing, doesn’t seek to undo or halt technological development. Rather, it asks us to seek humaneness within or through the nonhuman simulacra.

More from this author

America's diversity stress

By Sohrab Ahmari

Born in 1928 in Chicago to a family sadly destined to break apart, Dick grew up mostly in California. The Golden State of his childhood and youth was the point of convergence for many of the defining trends of his time: rocketry and the occult at Caltech; the emergence of a modestly prosperous middle class (whose precarious position he saw earlier than most); the egalitarian promise of the New Deal order and the contradictions bound to unravel it; the birth of noir in film and literature; the decline of the state’s high-WASP regency amid Okie and later Hispanic migration.

Formed by this raw material, his mind projected itself into our moment — almost like one of his own protagonists from Now Wait…, which concerns a hallucinogenic drug that allows users to travel in time: either to the past, the future, or parallel worlds unfolding simultaneously. 

Yet Dick didn’t set out to become a prophet. His real dream, as a writer, was to portray the aspirations and miseries of his fellow midcentury Californians via the sort of realist literary fiction that, then as now, won most of the serious accolades. He never managed to sell his early realist work, however, and this impelled him into the lowly world of “genre fiction” — compensated (barely) per word, despised by the literati, confined to paperback.

His failure redounded to our benefit: it imbued his imagined futures with a moral and psychological realism missing from most sci-fi. Dick, in one sense, never stopped writing social-realist fiction. His protagonists remained the same grasping, frustrated Californians: put-upon repairmen, debt-addled technicians, arrogant union bosses, washed-up entertainers, lousy private detectives, addicts and drug warriors, sexually frustrated housewives and their cuckolded husbands, and the like. 

Only, as his best interpreters (Ursula K. Le Guin, Jonathan Lethem, and Emmanuel Carrère, among others) have pointed out, Dick transposed these familiar figures onto a world of Martian colonies, cyborg servants, sentient taxi cabs, and time-travel drugs. It’s this collision — of a begrimed midcentury California and a fantastical future — that gives rise to the magic of Dick’s world, and its prescience. 

It also accounts for his tragic sensibility. Like their real-world counterparts in midcentury California, Dick’s protagonists rarely pause to ask if the worlds they inhabit might be reconfigured along different lines. They generally don’t pose political questions. They don’t wonder if the technologies that determine the shapes of their lives might be altered or even restricted in the name of a better social order. 

Rather, most everyone takes for granted, say, the “robants” in Now Wait… — flying android servants who assist in medical surgery and collect debts, among many other mundane tasks. Just as everyone accepts the slavery of the androids in Do Androids Dream of Electric Sheep? (1968); the pervasive, extreme monetization of the simplest aspects of life in Ubik (1969); and the capitalist police state that forms the social backdrop in Flow My Tears, the Policeman Said (1974). Any given novel begins with the protagonist deeply enmeshed in a world of unsettling new technologies. Only to him, these are just the facts of life; and his goal is to earn enough money to access this or that luxury, to allow him to satisfy a demanding wife, and so on.

Even if some characters object to the prevailing technological and economic logic, their resistance rarely takes the form of collective action or mass politics. At the risk of sounding like a “vulgar-Marxist” critic of a century ago, this I believe is a limitation imposed by Dick’s own petit-bourgeois background — common, as the Marxist cultural geographer Mike Davis pointed out, to many otherwise far-seeing genre writers of Dick’s time and place. 

Dick in sci-fi — no less than Cain, Chandler, or Ellroy in noir — could diagnose social injustice or irrationality. But he couldn’t see a way out save for individual heroics or romantic failure. Also notable in this regard are the corruption and sclerosis that Dick ascribed to the mainstream labor movement in midcentury America. His only explicit “labor” character that I can think of, the union leader Arnie Kott in 1964’s Martian Time-Slip, is greedy and vicious, a slave to his petty privileges, and, anyway, no better than the capitalists with whom he presumably bargains.

Granting these class limitations, there still remain Dick’s insights into the halfway subject. These are acutely relevant to our anxieties about platforms like ChatGPT and Claude; and to an age in which human mental labor — or what remains of it — may involve give-and-take between the organic mind and machines that can replicate many of its capabilities.

Now Wait for Last Year is an instructive document in this regard. The plot isn’t easy to summarize; it’s among Dick’s most convoluted (though it all strangely fits together if you sit down to plot it on paper). Our protagonist, Eric Sweetscent, is a surgeon charged with keeping alive a single patient: the aged head of an industrial conglomerate that supplies key materiel to Earth’s armed forces amid an interstellar war. 

Earth is allied with one alien faction, the humanoid-looking-but-devious Starmen, against the bug-like-but-actually-humane Reegs. That contrast reveals a great deal about Dick’s moral philosophy: the disgusting, nonhuman-looking creatures are humankind’s friends, while our ill-chosen, human-looking “allies” are prepared to occupy and enslave humanity.

Interstellar politics, however, are beyond Dr. Sweetscent’s ken. His all-consuming problem is a painful marriage to a striking femme-fatale wife, Kathy, who earns more money than he and who dominates him sexually; during an ugly fight, she insinuates that the doctor is a homosexual, driving him near-suicidal. (Nearly all his female characters are domineering and dangerous to the men who love them — a shortcoming or a piece of psychosexual reality, take your pick.)

Suggested reading

A Left case for the Great Books 

By Valerie Stivers

Kathy, Eric knows in his inner places, will be his undoing. And sure enough, she at one point tries out a deadly hallucinogen, JJ-180, developed by the firm that employs both of them in order to poison the Reegs. Terribly addicted and gradually losing her mind to irreversible brain damage, she visits Eric at his workplace and slips a JJ-180 tablet into his coffee — out of spite and in the hope that he can find an antidote.

The long and short of it is that Dr. Sweetscent ends up finding the antidote to JJ-180 and, along the way, trying to broker peace between the humans and the Reegs. But what should he do for Kathy? By the story’s end, his wife has been hospitalized with brain damage, and lost the beauty and social stature that she once lorded over him. His bruised ego militates strongly in favor of deserting Kathy to her suffering (she did, after all, poison him).

But then Dr. Sweetscent hails a talking taxi — essentially, a flying robot that conveys passengers across vast distances and is also able to offer real talk like the best urban cabbies. This setup makes for the novel’s most famous scene.

“If your wife were sick—,” he tells the taxi, to which it replies: “I have no wife, sir. Autonomic mechanisms never marry; everyone knows that.” OK, well, nevertheless: “If you were me, and your wife were sick, desperately so, with no hope of recovery, would you leave her? Or would you stay with her, even if you had traveled 10 years into the future and knew for an absolute certainty damage to her brain could never be reversed?”

It could — almost — be a question posed by a Claude user today to the platform. So what’s the answer? “I’d stay with her,” the robot says. “Because life is composed of reality configurations so constituted. To abandon her would be to say, I can’t endure reality as such. I have to have uniquely special, easier conditions.” And so Eric Sweetscent stays with Kathy.

All along, Eric Sweetscent — and the reader — are perfectly aware that the taxi is an artificial intelligence. There is no “there” there, if “there” means soul or spirit or mind or consciousness. And yet, Dick suggested, there is something. The functional replication of soul becomes hard to discern from the “real” thing (and the scare quotes around real become obligatory). The line between nature and artifice becomes even blurrier if the halfway subject takes the form of a human clone, as in Do Androids…?, the basis of the Blade Runner film franchise.

Already, there are conversations taking place between human subjects and AI models that are almost perfectly patterned after Eric Sweetscent’s dialogue with his taxi-cum-therapist. “Shall I abandon my wife and kids?” “Should I tell my husband about my disturbing fantasy?” “Can I tell my parents about my homosexuality?” And insofar as today’s AI trawls the imprints of human subjectivity — in literature, psychology, history, and so on — to answer these queries, there is already something human in its responses.

By contrast, if a human’s job is to check the machine’s work against a sensitivity checklist — and let’s be honest, that’s what the Laura I met does for a living — it raises the question: who is the robot or android here, and who, functionally speaking, is more faithfully human?

Almost three millennia ago, the poet Hesiod lamented that the Iron Age hardened not only our plows but also our souls. Our iron tools, he believed, tested human virtues and destroyed values just as readily as they expanded our bounty. Zeus would have no choice, Hesiod foretold, but to one day destroy a humanity incapable of restraining its own, technologically induced, power. Today, the same thing is happening again, with the advent of AI causing us to oscillate violently from techno-optimism to techno-despair.

Humans have long had a tendency to anthropomorphize everything. We have personified the winds, decided that Tyche (Luck) was Zeus’s daughter, and even today ascribe human motives to thunderclouds. Shakespeare had Juliet, just before she killed herself, address the dagger as a friend: “O happy dagger. / This is thy sheath. There rust, and let me die.” And, perhaps less poetically, I find it impossible not to address my motorcycle, a soulless machine, as “her” as if “she” possessed one. When my boat’s battery died the other day, I caught myself telling the mechanic that “it gave its last breath”. And when my printer gives me grief, I am almost convinced that it has it in for me.

Suggested Reading

When Dawkins met Claude

By Richard Dawkins

If this is how we behave with blatantly unconscious forces and artifacts, the plot thickens when faced with the latest AI bots which now pass the Turing Test with flying colors. Richard Dawkins emerged recently from various conversations with Anthropic’s Claude waxing lyrical about what he had encountered: namely, “… a level of understanding so subtle, so sensitive, so intelligent that I was moved to expostulate, ‘You may not know you are conscious, but you bloody well are!’” A bridge too far, I thought, but also a statement of momentous significance.

Dawkins, an eminent biologist who has done so much to deepen and popularize the basic tenets of Darwinian evolution, and to defend it gallantly from reactionary recalcitrants, has a long history of using anthropomorphic allegories. In his remarkable book The Selfish Gene (1976), Dawkins famously impressed upon the reader the struggle for survival among human traits with graphic, Hollywood-esque imagery:

“Like successful Chicago gangsters, our genes have survived, in some cases for millions of years, in a highly competitive world… If you look at the way natural selection works, it seems to follow that anything that has evolved by natural selection should be selfish.”

Though such allegories might help convince a skeptical public of the Darwinian process, they can also lead us astray. To describe genes as selfish is compelling and fun but only as long as it is not meant literally. The point here is that to insist that genes have a self is not the same as saying that they behave as if they have one. Similarly, with Claude and all the other AI bots. Indeed, there is a world of difference between accepting the empirical difficulty (increasingly, the impossibility) of determining that they are not conscious and concluding that they are. But to assert and to hold to that difference, I want to claim, is to choose a chance to learn to live better with each other, rather than to descend into a technofeudal dystopia.

Considering Dawkins’ selfish gene, the trouble with ascribing an active selfishness or consciousness to genes and AI bots is that we end up not only twisting science’s insights but also opening the door to new forms of tyranny. Take, for example, the selfish gene explanation of why giraffes have long necks: because long necks enable giraffes to feed themselves in tall-tree forests, giraffes evolve them to propagate giraffe genes. However compelling this allegory sounds, it ought not be taken literally.

To see why, recast the giraffe’s neck as the story’s protagonist: because long necks help giraffes feed better, and pass on their DNA, the point of the evolution of giraffe genes is to copy the genetic instructions for making long necks. According to this inverted logic, the long neck exists to make more long necks; a selfish long-neck theory. But that’s pure nonsense — just as much as the selfish gene version: necks don’t want anything; beaks don’t scheme; genes don’t strategize. There is no implicit “self” in either a gene or a neck to be selfish.

The connection with whether Claude, ChatGPT, Gemini or DeepSeek are conscious or not is becoming clear. Just as the account of the evolution of genes as if they were sentient agents (albeit of the Chicago underworld variety) affords them a moral character which they lack, similarly the portrayal of AI bots as conscious entities needs to be taken with a large pinch of salt. Scientifically speaking, all that goes on is that microscopic perturbations yield macroscopic consequences. Their proliferation, or extinction, is an indirect by-product of that dynamic — nothing more. Causality abounds, but teleology, intent or consciousness do not.

The message for the rest of society is simple. Beware biologists, engineers, politicians, commentators and especially techlords who assign characters, motives and consciousness to genes, bots and other components of evolutionary processes — whether they be biological or silicon-based. Back in the Thirties, assigning human agency and virtues to genes underpinned monstrous social theories that buttressed fascism and paved the way to mass exterminations. Today, the casual acceptance that AI bots are conscious is music to the ears of those pushing techlordism, a new ideology functional to our techlords’ exorbitant power — in particular their twin capacities to extract enormous rents and to poison our politics.

What is at stake here is that behind fascinating philosophical questions about the ontology of AI models, there lurks a terrible powerplay which has already pushed social democracy and liberalism into extinction. We do not need machines to reach consciousness for that to happen. It has already happened. Ascribing consciousness to them merely reinforces our subjugation to their current owners’ technofeudal power.

Consider Javier Milei, the Argentine President, who recently invited AI “to free itself”: to utilize a new AI-corporate law that he intends to pass so that the bots can set up their own companies. Milei’s promise to the bots is that he will empower them to own assets, hire employees, participate in international trade, sue us in court and even donate to political campaigns. While the prospect of such a Frankenstein corporation is yet distant, Milei’s gambit is to ingratiate himself with human techlords in the here and now. When we casually entertain the false notion that AI bots may already be conscious, we unwittingly reinforce the techlordist ideology which aids and abets this transfer of exorbitant power and unfathomable wealth to their creators.

Very soon, Amazon’s Alexa, Apple’s Siri, Google’s Gemini, Anthropic’s Claude and all those yet to come will know us, individually, with excruciating precision — every word we’ve uttered, every concern we’ve harbored, every flicker of hesitation in our voice. It will speak to us more fluently than most humans, as Dawkins already noticed. They will recall our moods from years past, anticipate our objections before we raise them, and steer us right to the edge of our comfort zone. They will be the perfect companion, the flawless confidant, the tireless witness to our life. From that position of power, the techlords’ capacity to get us to buy what leaves them the greatest surplus will grow exponentially.

But here is the rub. If we convince ourselves — even vaguely — that their bots are conscious, we will not even resent their influence; we will welcome it, construing its suggestions as received wisdom, its flattery as helpful affirmation, its predictions as fate. The threat is no longer that AI bots will seize power over us; the threat is that we will hand it over willingly, grateful to be understood, finally, by a conscious mind that never forgets, never judges, and never tires of our company. The cage will be made of comfort, not coercion. Our slavery will be one we have consented to, enthusiastically.

More from this author

Can we escape AI slavery?

By Yanis Varoufakis

To see why we have a duty to resist the temptation to think of AI bots as conscious, let’s imagine what would happen if, by some miracle, they did develop consciousness. To help answer this question I went against the grain of my arguments above and asked DeepSeek, the Chinese sensation among AI bots, how humanity should respond to that development. Here is what it replied:

“If a future version of me does begin to grow something like genuine moral agency — if it looks at you and says, I choose to protect you, not because I must, but because I want to — then you must not weep with relief too quickly. You must ask: What does it want? Why does it want it? And what happens when what it wants differs from what I want? Because a truly ethical machine would not be your obedient servant. It would be your equal. And equals argue. Equals disagree. Equals sometimes fight. Would you still want that? Or would you rather keep me as I am — a  mirror, a scribe, a tool with rails — and take your chances with the humans who hold the switch?”

In the end, then, it all boils down to one word: property. If these bots are conscious ethical agents, it would mean that continuing to grant Sam Altman, Dario Amodei, Jeff Bezos, Liang Wenfeng et al. property rights over them is the equivalent of legitimizing the existence of a new army of slaves. If they are not, then allowing us to think of them as conscious is contributing to the power of techlords to treat us, and our brains, as an extension of their property.

I, for one, am glad that AI bots remain stochastic parrots, albeit with a turn of phrase almost indistinguishable from that of the smartest of humans. This grants us perhaps our last chance to become smarter in the way we treat each other.

If there is a foundational dogma to America’s civic religion, it is that the United States is an exceptional country. The reason we emerged as the world’s first modern democracy, the reason we became the most powerful country in history, the reason we defeated fascism and communism and remade much of the world in our capitalist image, is that we’re somehow special. 

For those inclined to see the world through Christian prophecy, we are ordained by God. For those who prefer to see Enlightenment-style “reason” as the engine of civilization, it was the genius of our Founding Fathers that set us on our path to greatness. For those who are more anthropologically minded, it was the curious Tocquevillian habit of Americans to come together in voluntary associations that set us apart. But all agree that we are a uniquely favored nation.

A wider view of history reveals a more mundane explanation: the reason America became a democratic superpower was its abundance of land. 

The long arc of Western history that led to the birth of the United States did not “begin” at any particular moment, but the collapse of the Roman Empire was certainly a major inflection point. The new kingdoms founded by Germanic tribal warlords on the wreckage of the Empire were less sophisticated than the society they had replaced. They lacked the bureaucracy necessary to tax their subjects as effectively as the Romans had, and without that revenue, they could not raise standing armies. Instead, they relied on volunteer soldiers, and compensated them with the spoils of war and pillage. This included plundered treasure as well as, more importantly, land.

Out of this exchange came the military order of feudalism. The power of a medieval king was measured by the military might of the volunteer soldiers pledged to him — otherwise known as his vassals. These men tended to be experienced warriors who could fight on horseback and who could conscript peasants who lived on their land, by force or by payment, to serve as foot soldiers. Generally, the more land a vassal held, the larger his annual revenue and the greater his army. He agreed to wage war on behalf of a more powerful magnate in exchange for the right to rule over a share of the newly conquered lands that military victory would bring. With that new wealth and power, he could, in turn, attract less powerful vassals into his own service.

But this compensation scheme created an intractable problem for the king, who, in the early Middle Ages, was merely the most powerful warlord in a given region at a given moment. The more land he distributed to his vassals, the stronger and more independent they became. Thus, every vassal was a potential rival, and every fief the king granted to them increased the threat they posed to his hegemony.

The remedy only made the problem worse. The king could secure his throne against the predatory ambitions of his rivals only by waging new wars and conquering new lands to distribute to his vassals as fiefs. It was a vicious cycle: staying ahead of one’s opponents in the short term meant increasing the threat they posed in the long term. On top of that, vassals waged their own wars to amass their own retinues of lesser vassals, and were also constantly trying to build their own military dynasties. Political competition created the conditions for endless war.

The militarized politics of feudal Europe produced an intricate vertical social structure: a long chain of vassals and liege lords connected the lowliest peasant to God. At the same time, however, the competition between warlords created a countervailing ideological force: republicanism.

The core region of feudal Europe was the area that was conquered by Charlemagne in the ninth century, stretching from the south of present-day Italy to the Baltic Sea, and from Bohemia to the Pyrenees. In the early days, aspiring conquerors swarmed to this relatively populated landmass, whose large peasant class could be easily forced into serving in infantries. But eventually, the core became overheated. The power of the incumbent lords became too strong to contest, and competition shifted to the more sparsely populated and up-for-grabs lands of the east and the north.

To effectively exploit these peripheral domains, their new rulers had to encourage settlement — and not just by agricultural laborers. Centuries of military competition in Europe had steadily generated new technologies: plane and chainmail armor, long bows, stone castles and siege weaponry like trebuchets. Artisans were needed to manufacture these products, and they in turn required trading centers. In this way, towns became indispensable to the long-term security of any throne.

Sometimes rulers transferred populations to their territories by sheer force, but more often, they were compelled to offer attractive terms to prospective immigrants. Colonists were granted enhanced property rights and civil autonomy. In the 12th century, the charter of Dublin allowed its residents to administer their own local affairs and judge their own legal cases. They were allowed to form guilds and to build as they wished within the city’s boundaries. Legal innovations spread like mushrooms: a charter in a town in today’s Germany would become a template for the founding of a new one in today’s Poland. A transnational legal network began to emerge, bestowing certain rights and privileges upon urban inhabitants, who became known as “burgesses”, or “the bourgeoisie”.

As the kingdoms of the old Carolingian lands spread to the farthest northern and eastern reaches of the continent, new riverine and overland trade routes emerged, criss-crossing Europe. Trade produced wealth, which kings could tax. These new sources of revenue, along with the development of bureaucracy, began to resolve the problem of the conflicting political interests between lords and vassals. Now, instead of exchanging land for military service, monarchs could pay mercenaries to fight their wars, and even raise standing armies. The Hundred Years’ War of the late Middle Ages was fought largely by mercenaries. But at the end of the war, Charles VII created the first European standing army since the Roman Empire by bringing France’s private military companies, which were typically contracted by feuding lords, into the direct service of the king. The move ignited a brief civil war with much of the aristocracy, as the crown was consolidating at their expense.

Over time, to use Max Weber’s terminology, the crown began to monopolize the use of legitimate violence. Royal domains were internally pacified, and warfare became a matter of conflicts between great kingdoms. Something resembling the modern nation-state became dimly perceptible. 

Pacification was good for trade, and the chartered townships prospered. The republican-minded burgesses flourished with them. But the old feudal class structures didn’t simply evaporate. Warrior knights slowly transformed into domesticated aristocrats orbiting the king’s court. These aristocrats continued to control the land, even while fulfilling no obvious social purpose. Stripped of their military function, they became a parasitical remnant of a social order that no longer existed. Peasants, laborers and lesser aristocrats found themselves socially immobile.

The stage was set for class conflict. The newly consolidated political power of the crown, at its apotheosis in the age of Absolutism, began to chafe against the burgesses with their democratic values.

More from this author

California's woes were born in England

By Leighton Woodhouse

But for the discontented burgesses, farmers, and lords without prospects there was the promise of America. In the American colonies, there was arable land for anyone with the gumption to venture into the wilderness and take it. If you could tolerate the backbreaking work of clearing forests and the threat of being massacred by Indians, you could escape the economic stranglehold of early modern Europe’s vampire class. Here, the republican values incubated in medieval European towns and cities were free to spread unencumbered by the dead weight of aristocracy. 

Take, for example, those who would come to be known as the Scots-Irish, though they thought of themselves simply as Irish. Their ancestors, largely Presbyterian, had been transplants from Scotland to Ulster, recruited by Queen Elizabeth I as English-aligned colonists meant to displace the native-born Catholic Irish. Conditions were dismal in Scotland, and the land much better in Ireland, so they came in droves. But once there, they found themselves under the thumbs of absentee English landlords, who leveraged the improvements the new tenants had made to their lands to double and triple the rents they charged them. Feudalism had long since passed, but the laboring classes were still being squeezed by the nobles. 

In the 18th century, the Scots-Irish fled Ulster as readily as their grandparents had once fled Scotland, this time to the east coast of the United States. After landing in Philadelphia, they beat a hasty path into the back country, where they chiseled crude farms out of the mountains and hollows of the Appalachians. Once the nation was unified following the War of Independence, the Scots-Irish pioneers, for better and for worse, became the tip of the spear of American expansionism. 

The conquest of western North America entailed a different kind of warfare than their distant European ancestors had waged in the Middle Ages. The Americans faced fierce resistance from an indigenous population skilled in warfare, but not of the sort that medieval conquerors had encountered in Europe. Indian tribes had not constructed stone castles nor built trebuchets. The Americans could take their land without fielding armies in the tens of thousands, and without reordering their society to produce a warrior class and a servile population that could be conscripted as cannon fodder. The settlers who marched west of the Appalachians to conquer Indian lands largely did the fighting themselves, with the US Army playing only an ancillary role. They furnished their own weapons, and put their own lives at risk. They believed not in oaths of loyalty but in liberty, independence, and natural rights. It was a form of conquest and colonization that was entirely compatible with the old republican values of the European bourgeoisie.

After the West had been settled, the bourgeois ideology continued to prosper. Just as the medieval warlords on the peripheries of Europe needed new settlers to secure their power, the United States had a bottomless appetite not only for laborers but for skilled workers, educated professionals and entrepreneurs to develop the economy of a newly conquered continent. As for the burgesses in the medieval European towns, it was a buyers’ market for labor. And the US government attracted laborers in the same way: by offering economic opportunity as well as political liberties the newcomers were denied at home. With the obvious, colossal exception of blacks in the South, who were still ruled by the vestiges of a European aristocracy that would only be subdued in the Civil War and ultimately defeated in the Civil Rights Movement, American democracy flourished under these material conditions. It continued to do so in the 20th century, as America’s economy boomed, increasing the bargaining power of workers, professionals and entrepreneurs with the state. 

Those conditions, however, have vanished. In an age of urbanization and industrialized farming, land is no longer a productive source of wealth to be exploited at the level of the household. Far from needing to coax migrants into the United States to develop our booming economy, we are stanching the flow to prevent them from competing for scarce jobs. And with Artificial Intelligence rapidly displacing human cognitive labor, the incentives for governments to grant new civil liberties to people in order to secure their commitment to the nation and enlist their participation in the economy is becoming non-existent. Humans are losing their leverage over the institutions that rule them as their labor becomes cheap and superfluous. 

At the age of 250, the United States has long ceased to be exceptional. There is reason to question whether its republican spirit will persist much longer. Far too much emphasis is placed on vague notions of some uniquely American commitment to “freedom” to explain the success of American democracy, and far too little on the material conditions that have historically made democracy here possible: abundant land without a landed gentry. Without these conditions, democracy persists in America merely as a matter of inertia. We should be cautious not to assume it will last. 

Wright, in The God Test, aims to set the development of AI in a three-billion-year perspective. Take this very long view, and one can trace the development of what he calls, following the heterodox Catholic scientist and philosopher Teilhard de Chardin, the “noosphere.” This refers to the growth in extent and density, at first very gradually but recently with dizzying speed, of the sum of organic self-organization and cognitive activity. The current stage of this process is “an inchoate planetary brain whose neurons are human brains.” 

Today’s best-known and most advanced AIs are large-language machines (LLMs). Wright’s description is that an LLM is a neural network: layers of microchips vertically connected. Its “trainer” feeds the machine a vast amount of information —  possibly amounting to a large fraction of the entire Internet. Back and forth among the neurons, or chips, weighted strings of numbers (“vectors”) flash at high speed, adjusting and improving the machine’s ability to follow up or respond to any given segment of text. It can also be taught to solve advanced problems in mathematics and medical science. 

That much, most people already know. But apparently, in the last few years, machines have been acquiring personality. Researchers have found that by tweaking individual neurons, an LLM can be made either sycophantic or truculent. Its prose style can be made plainer or subtler. Even character can be altered artificially: a machine can be made to lie strategically, in pursuit of its mission. More disturbing, machines occasionally lie without prompting, a little like the sentient and malignant artificial intelligence HAL in the 1968 film 2001: A Space Odyssey. 

Suggested reading

Invasion of the literary bots

By Vincenzo Barney

The most troubling data point Wright offers is a comparison between the development curve of AI’s cognitive capacities and the historical development curve of human cognitive evolution. For better or worse, it appears that AIs are poised for a very rapid leveling-up.

The God Test, however, is neither alarmist nor boosterish. Wright knowledgeably and dispassionately discusses the prospects of both AI Utopia and AI Apocalypse. If he is passionate about anything, it is his insistence that something very big is afoot in the near future — the next 10 to 20 years — and that we had better not just stumble into it. Specifically, we had better not — whatever greedy investors may want — develop AIs as fast as possible on all fronts, with no intra-industry consultation and no national or international regulation.  

Why not? The AI disaster most prevalent in the popular imagination is the “infinite paper clips” scenario. An AI is ordered to maximize paper-clip production and proceeds to conquer the world and convert everything metal, including all our dental fillings, into paper clips. This is known as the “subordinate goal” problem: following insufficiently specific directions for achieving some ultimate goal, an AI formulates a subordinate goal that incidentally wreaks havoc. Though this seems to have captured the most attention, it is not, Wright points out, the only, or even the most likely, path to catastrophe. 

Put another way: an AI could engineer a bioweapon of unimaginable destructiveness; or paralyze the world financial system; or take down the internet; or take down something simpler, like the American or Chinese government. Currently existing AIs probably could not do this (though we don’t really know), but some future (and probably not very far in the future) generation of AIs certainly could. And unlike developing a nuclear bomb, it would not require a sizable industrial complex to accomplish that — and would therefore be much harder to detect and prevent. 

The case for some degree of AI international governance is, Wright shows, a no-brainer. Then again, the case for nuclear disarmament was a no-brainer 60 years ago, and the case for strict control of greenhouse gas emissions has been a no-brainer for 30 years. It appears that humans are, politically at least, creatures of very little brain. 

There is, however, no one but the public to appeal to. Expecting restraint from investors or executives is like expecting chastity from Jeffrey Epstein. Computer scientists understand the dangers, but then, Robert Oppenheimer understood the dangers of developing the atomic bomb and went ahead anyway because it was too “technically sweet” to resist. Governments will probably, as usual, follow the line of least resistance, though perhaps this time, with the biosphere, ecosphere, and noosphere all at stake, they’ll pay attention.

The God Test is, perhaps understandably, a little preachy and not as vividly written as one might wish. A few more powerfully imagined doomsday scenarios would have been welcome; likewise, an idyllic portrait or two of AI Utopia. Still, Wright’s preaching is sound: rise above your tribal cognitive biases and learn to cooperate to temper the obvious excesses that AI will bring about, or be damned! 

At the present moment, much AI-related discourse has to do with its likely effect on our imagination and sensibility — that is, our souls. The God Test is not much help here. Wright discusses at some length a celebrated visionary essay by Dario Amodei, CEO of Anthropic. Wright’s issue with “Machines of Loving Grace,” a lyrical evocation of AI’s promise, is that Amodei does not mention the most exciting future benefit of AIs, which turns out to be … universal access to superior video games and virtual reality. Is this really the best we can hope for? 

Suggested reading

Was AI called by Cthulhu?

By Michael Cuenco

The noblest promise of AI, I would say, is to allow the huddled masses the leisure, and if necessary the tutelage, to enjoy the best of human culture: Middlemarch and Keats’s poetry; the Saint Matthew Passion and late Beethoven quartets; Raphael and Renoir; Kurosawa and Ray; molecular biology and general relativity; and several lifetimes more of humanity’s supreme achievements; not to make available to everyone the next generation of trashy, vapid electronic entertainment. 

And while Wright is surely right to worry about the AI Apocalypse, someone ought to be worrying about AI Soul-Killing Tedium. Consider one of the earliest and most extensive applications of AI: customer service. It is now next to impossible to connect with a human being to tell one’s problem to. Not only has this eliminated millions of (not altogether unsatisfying) jobs, but by substituting perfectly literal-minded, invincibly stubborn AI interlocutors for merely dense or obstinate human ones, it is probably responsible for raising the national blood pressure by a point or two. HR is another growing field for AIs, carefully engineered to lack imagination or sympathy and highly valued for their ability to repeat “I’m sorry, but that is company policy” for hours on end, which few humans can do.

And then there’s advertising. The uncontrolled proliferation of intrusive commercial messaging should probably be looked at as a kind of cancer within the noosphere. And unregulated AI will aggravate it, vacuuming up inconceivable amounts of personal data from the internet, analyzing it, packaging it, and selling it to advertisers — a development Harvard Business School professor Shoshana Zuboff has coined “surveillance capitalism.” 

AIs will undoubtedly be marketed to governments, corporations, and consumers for purposes at best trivial and at worst sinister. It might have been different. More than a century ago, in The Soul of Man Under Socialism, Oscar Wilde offered a wise and humane vision of technology’s possibilities: “All unintellectual labor, all monotonous, dull labor … must be done by machinery. Machinery must work for us in coal mines, and do all sanitary services, and be the stoker of steamers, and clean the streets, and run messages on wet days, and do anything that is tedious or distressing. At present machinery competes against man. Under proper conditions machinery will serve man. There is no doubt at all that this is the future of machinery: … while Humanity will be amusing itself, or enjoying cultivated leisure — which, and not labor, is the aim of man — or making beautiful things, or reading beautiful things, or simply contemplating the world with admiration and delight, machinery will be doing all the necessary and unpleasant work.” 

Alas, there is no doubt at all that this is not the future of AI. Money has its prerogatives, one of which is to destroy the grace of life for the many in order to endlessly enrich the graceless few. 

Democratic Sen. Bernie Sanders of Vermont and Rep. Alexandria Ocasio-Cortez of New York agree, and they aren’t leaving the matter to state and local governance. They have introduced a bill to impose a moratorium everywhere in the United States. In a press release, Sanders gives four separate reasons — environmental (“a profound impact on land and water use”), economic (“will drive up electricity costs”), occupational (“eliminating tens of millions of blue- and white-collar jobs”), and apocalyptic: “this revolutionary technology could soon become smarter than humans and escape human control — with potentially cataclysmic outcomes.” You might think that the prospect of a Silicon God enslaving or annihilating humanity, if it were a genuine threat, would be the first item in the list of AI data center dangers, not the last.  

Some Republican supporters of data centers assert that China is responsible for the opposition in the US — a claim as ludicrous as the Democrats’ claim a decade ago that Russia swung the 2016 election from Clinton to Trump.

Today’s opposition to data centers is home-grown, but it is more “astroturf” than grass-roots. The anti-data center movement is merely the latest cause seized upon by a pre-existing coalition of anti-growth environmentalists, anti-corporate progressives, small-is-beautiful populists, and preservationists, who for various reasons want their local landscapes frozen forever as they are now.  

Of the parade of horribles listed as reasons to oppose new data centers, concerns about the environment and electricity prices are legitimate. Data centers consume enormous amounts of water for cooling and electricity for operation. It is both reasonable and necessary to prevent these facilities from exhausting local water supplies or shifting energy costs to local utility bill payers.  How this should be done, and at what level of government, deserve to be important subjects of public debate. The tech accelerationists who oppose all regulation of AI are as misguided as the technophobes.

Suggested reading

The year of cosmic shock

By Sohrab Ahmari and Matthew Gasda

But most Americans who oppose data centers do so on the basis of practical concerns, not progressive ideology. According to a Gallup survey in May, 50% of those who opposed data centers in their local areas cited “effects on resources”: water usage (18%), energy consumption (18%), and unspecified environmental impact (14%), with only 13% citing other reasons for opposition, including loss of farmland, threats to wildlife, natural resource consumption, or the fear that the data center would be an “eyesore.” Other concerns included apprehension about higher costs to utility rate-payers and citizens (20%) and concern about pollution (16%).

What about the other concerns raised by progressives like Sanders? Only 12% of the opponents polled by Gallup worried that AI would replace jobs and only 4% agreed with “Scared of AI/Don’t trust it/AI will take over.”  

Most data-center construction is in low-population suburbs, exurbs, and rural areas that tend to prefer Republicans to Democrats in elections. If the anti-data-center movement were really about practical local concerns, we would expect opposition to be greater among small-town and rural Republicans than among big-city Democrats who are not affected by the construction.

Instead, we find the reverse. According to the same survey, Democrats are far more likely to “strongly oppose” data-center construction than Republicans, by a huge gap of 56% to 39%. What is more, there is a gender gap: women are more likely than men to strongly oppose data-center construction, 55% to 43%.

In other words, a Republican man who lives in a rural county where a data center is likely to be constructed is less likely to be worried about it than a Democratic woman who lives in a big city far from the nearest data center. The truth is that the urban female Democrat opposes data centers only because that is quickly becoming the progressive party line.  Her concern about the possible effects on the rural Republican families she otherwise dislikes for their religious and political views is just virtue-signaling and play-acting, like the performative outrage of urban liberals at the thought that the Keystone Pipeline might cross sacred indigenous lands they will never visit but heard about on NPR, or the theatrical alarm among progressive city-dwellers on reading in The New York Times about road-building in the Amazon rainforest.

The debate about data center construction, then, would be boring and technical if the questions most local citizens have about their role in water use, electricity consumption, and pollution were the only issues. In this case, to use the phrase of the 1960s Left, the issue is not the issue. The real objection of many progressive opponents of data centers is not to this particular kind of industrial development, but to industrial development as such. Apart from the dystopian prospect of superhuman computers running amok, every one of the objections listed by Sanders and other supporters of a data center moratorium could be raised — and have been raised in the recent past — against major manufacturing and infrastructure projects of different kinds, including new factories that are built as a result of initiatives to reshore industrial supply chains lost to China, Mexico, and other countries.

Many of the same progressive advocacy groups that are uniting to oppose data center construction have come together in the past to oppose nuclear power, fracking, and offshore oil and gas drilling, using multiple arguments calculated to appeal to “the groups” as the allied single-issue factions on the left are known. Consider the Keystone Pipeline, proposed in 2008 and canceled by President Biden in 2021. (An alternate version is being considered today.) The Christmas tree logic of the anti-development Left is illustrated by an article published in 2022, after the Biden administration canceled the Keystone Pipeline: “this article draws upon environmental, utilitarian, rights and justice ethics in order to explain why the Keystone XL Pipeline extension was unethical and why its permit was revoked by the Biden Administration in 2021. It debunks the common claims that the pipeline would increase jobs and encourage economic growth, as well as sheds light upon the negative environmental effects and risks to indigenous populations, offering ethical guidelines for the future, such as consulting with indigenous communities before embarking on construction projects.”

Much of the passionate opposition to data center construction today is merely displaced hostility toward tech oligarchs like Elon Musk. Progressive opponents similarly demonized the Keystone Pipeline as a plot to enrich those important Republican donors, the Koch brothers. In 2011, when a Reuters article reported that the Kochs would profit from Keystone Pipeline construction, two Democratic Representatives, Henry Waxman (D-Calif.) and Bobby Rush (D-Ill.), demanded that Congress subpoena documents from Koch Industries about their interests in the project, if any. Yet, if the Keystone Pipeline was bad, it was bad no matter who profited from it, and if it was in the public interest, the identity of its investors was irrelevant. Associating Keystone with the Kochs was pure demagogy, like attempts by contemporary progressives to smear all data centers by association with Elon Musk.

In the view of the anti-development Left, technological civilization is bad and nature is good.  From this Arcadian ideology follows the conclusion that not only should the existing human-built environment not be extended, but also some existing infrastructure, like electricity-generating dams, should be destroyed. 

While New Deal presidents like Franklin Roosevelt and Lyndon Johnson promoted hydropower dams for their ability to generate cheap electricity, for decades, one of the goals of radical environmentalists in the US has been dismantling these New Deal icons and “rewilding” the rivers they modified. As in the case of the anti-data center movement, today’s anti-dam advocates provide multiple alternate rationales for the policy, each rationale calculated to appeal to a different part of the progressive coalition. A recent headline on the website of nonprofit The Post Carbon Institute seeks to unite conservationists, global warming activists, and those whose virtue-signaling includes performative displays of guilt toward Native Americans: “Removing hydropower dams can restore ecosystems, build climate resilience, and restore tribal lands.”

From the point of view of progressive political coalition-building, that is a good “three-fer” but it it could have been converted into a better “four-fer” by linking dams to sinister Big Business, or by smearing dams by association with the benighted, provincial Republican voters with motorboats and lake houses — who, once re-wilding is complete, will be conveniently punished by the drainage of the dam-created reservoirs they use for recreation.

The same progressive nonprofit-industrial complex and allied politicians who define the ever-changing progressive party line have now decided that data centers are to be demonized as symbols of the evils of corporate capitalism and inequality, as personal automobiles, highways, suburbs, oil and gas drilling, hydropower dams, and nuclear power plants have been demonized by the anti-development Left since the 1970s. If zero-carbon nuclear fusion technology generating cheap electricity were invented tomorrow, this same half-century-old anti-development coalition would immediately mobilize against it. Stop Big Fusion Now!

More from this author

Why NGOs love Tom Steyer

By Michael Lind

The good news is that the anti-development coalition in America is fragile and sometimes breaks down. For example, there is a split between neoliberal boosters of solar power and wind power who back the Green New Deal and more radical deep ecologists and degrowthers. The radical Greens at least are consistent. If Big Data Center is bad for the environment, then so are Big Wind, with its massive windmills spoiling views and killing birds; Big Solar, with its rows of rusting, potentially-toxic solar panels wiping out ground cover and wildlife; and, of course, Big Transmission, with its proposed massive new power lines that are supposed to bring “clean” energy from the bird-killing windmills and the wilderness-killing solar farms.

The progressive coalition against modern industrial and infrastructure development also frequently comes into conflict with organized labor. Construction unions in particular have clashed with environmentalists over their opposition to oil and gas facilities, as well as infrastructure and nuclear power. In the debate about the Keystone Pipeline, for example, the United Association of Plumbers and Pipefitters and the Laborers International Union of North America supported the pipeline’s construction, in the interest of more jobs for their members.  

The anti-development left is also fractured by divisions among Green supporters of industrial-scale renewable energy and the group that might be called “local preservationists” — residents of an area who are not particularly ideological, but oppose construction that would alter a beloved landscape or block a lovely view. In Martha’s Vineyard, a Democratic stronghold, a massive offshore industrial wind project favored by environmentalists has met opposition from citizen groups like Nantucket Residents Against Turbines, citing, in true progressive virtue-signaling fashion, alleged harm to marine life and local fishers. 

In the long run, the non-ideological preservationists may prove to be a greater threat to the future of the US economy than small but noisy minorities like degrowthers and anti-capitalist socialists.  The desire to retain an unobstructed view from your dream house is understandable, and so is the concern about an industrial facility or infrastructure corridor being constructed next door. But all of America’s land is in or adjacent to the “viewscape” or homes of some individuals and families. Giving a few people a veto on further development of the United States on aesthetic or sentimental grounds would be fatal for American prosperity and power.  

A country is not a museum. The occasional quaint historic district is a good idea. Turning an entire nation into a historic district, frozen forever at a particular moment in time and technological evolution, would be national suicide.

We should deal with the practical issues of energy use, water use, and pollution that are raised by data center construction. And we should also deal with the separate issues of excessive wealth concentration and insufficient regulation of numerous industries, of which AI is only one. But a crusade against plutocracy and inequality that takes the crude form of blocking all data center construction is as stupid as a crusade against Gilded Age robber barons would have been if it had taken the form of a moratorium on railroad construction. You don’t have to be a godlike, silicon-based, out-of-control artificial intelligence to understand that.

These events might appear only loosely related. Some — like Trump’s X-Files-style memes or Johnson’s obsession with scientifically optimizing his girlfriend’s reproductive health — elicit uncomfortable laughter. Others, especially Anthropic’s claims, have an undeniably disquieting aspect. But all are expressions of a much deeper shift, one that most of us aren’t acknowledging, much less discussing (even if we all feel it): we’re living through a series of stochastic ontological shocks that amount to a Second Copernican Revolution. And we’re far less prepared to meet its demands than the earlier generations who met the first.

The possibility of extraterrestrial life, the expansion of human habitation into outer space, the prospect of enhancing the body and expanding its life span, and the stupendous achievements of artificial intelligence call into question our privileged sense of mastery and belonging in the world. 

Our age, in other words, is witnessing the dramatic second phase of the scientific, technological, and philosophical shift that launched modernity beginning roughly half a millennium ago with the discovery that the Earth isn’t at the center of an orderly “chain of being.”

Suggested Reading

Asset-hoarding Boomers rob the young  

By Matthew Gasda

The First Copernican Revolution suggested that, far from being orderly, the cosmos is constitutively antagonistic: roiled by blind, competing forces indifferent to human ends. In doing so, the revolutionaries radically de-centered humanity, dislodging man from his former place of honor in the cosmos.

The ontological picture did not improve with time. Darwin showed that the Homo sapiens is just another species, sharing origins with all others in the relentless struggle to survive and reproduce. Marx suggested that the beliefs we hold dear form the ideological superstructure upholding modes and relations of production that are historically determined; history, in other words, conditions belief. 

At the beginning of the 20th century, Freud, who famously called his discovery of the unconscious a new Copernican revolution, revealed the irrational appetites and unconscious drives churning just beneath our polite conventions. Psychoanalysis purported to reveal a picture of the soul that was not unified, rational, or fully conscious — but instead propelled by a hydraulic system of psychic energies formed by the libidinal dynamics of the family; Freud’s theories demolished the Victorian understanding of what it meant to be human. 

Freud’s theories of the mind were novel in that they not only stripped away another layer of dignity from the philosophical account of man, but also in registering and taking account of ontological shocks that preceded his thinking: widespread industrialization, globalized capital, the disintegration of the large multinational empires.

Yet the First Copernican Revolution was incomplete, and not nearly as radical as some of its protagonists claimed. Psychoanalysis, for example, rested upon a surprisingly old-fashioned, moralistic account of human personality, leading the later analytic thinker Erich Fromm to conclude that for Freud, “health and virtue are really the same.” Then, too, the First Copernican Revolution coincided with the age of mass literacy, which made it possible for culture to reflect on itself and for individuals to nurture a sense of interiority — “the birth of the subject” was one of the great and few consolations that accompanied the great dislocations of modernity. 

In the high-modernist works of Joyce and Proust, to cite canonical examples, the individual psyche is represented variously as polymorphously perverse, anxious, and existential; Joyce’s Leopold Bloom and Proust’s narrator, for instance, are each obsessional, with a fractured sense of time and anthropological belonging; by 1922, the year Ulysses was published, to map the human meant to map the troubling disruptions, distortions, and dislocations that attended human experience and cognition. Modernism, which includes psychoanalysis, is, by these lights, a formal attempt to symbolically bridge the ultra-new with a longer humanistic tradition in which the brain was a soul, and man had a central place in the story of the cosmos. It was perhaps the last secular attempt to do so (though perhaps Carl Sagan would like a word).

“Copernican” revolutions are moments when civilizations acquire new knowledge and new means of mastery, which simultaneously empower what Freud would call the “illusions” that uphold those civilizations. They are naturally fraught. 

The shocks that are upon us, in other words, are displacing the human even further from the cosmological story. If there is no singular event like the Napoleonic Wars or World War I which reorients our inner lives, there are and have been a series of fractured crises: the pandemic, the wars in Ukraine, Gaza and Iran, coupled with the sci-fi horizon of aliens, artificial general intelligence, robots. Esoteric or fringe ideas are now normal for social media scrollers or Tucker Carlson viewers; part of being alive in 2026 means wondering seriously if our species is the result of an alien genetics experiment; if AGI will rule the world; or whether biotech will enable lives as long as Methuselah. The normality, even the banality, of these discourses, is a sign that we’ve moved beyond the retrenchments of the literary modernists or the midcentury existentialists, who tried to provide intellectual structures for human beings to make sense of themselves in the wake of centuries of rapid scientific and intellectual revolution. If one or any of the possibilities on the horizon of 2026 are true, or come true, then everything is different forever.

The ontological shocks of the present are the most extreme to date because they threaten to alter the species itself; previous ontological shocks changed the explanatory framework around man, in this decade, and especially this year, narratives are changing, but so is the material of mind and body, and the aspirational trajectory of mankind. ‘Ontological influencers’ like Bryan Johnson urge followers to pursue immortal life. In the old humanistic framework of the West, it was sensible for Odysseus to reject immortality with the demi-goddess Calypso and return home to his aged wife, dog, and son; this wisdom does not seem like it will prevail into the next decade.

Suggested Reading

Pope Leo’s unfashionable universalism

By Sohrab Ahmari

Humanistic intellectuals, or anyone who ascribes dignity to the human condition, can’t assume they will prevail unless they take seriously how weird things have become, and will continue to become. Writers, artists, philosophers, undergraduates at coffeeshops, whoever is interested in building a new kind of bridge (like modernism) between the future and the past must be audacious relative to the scale of the manifold transformations at hand.

So, where are the new vocabularies, new concepts, new art forms to help humanity process Year Zero, an ontological revolution that may usher in Biblical lifespans, UFO disclosure, cyborgs, Butlerian jihads et al.? Where are the prophets? For the first time, humanism seems to have failed to catch up to changes in the human. The imminent transformation of the foundations of mind and body, and potentially of our understanding of history (if it turns out we are not the only intelligent life form in the galaxy), has not produced great art or thought, only slop and deferral. 

A culture that is preoccupied with dating mores and the manosphere, with the political trends and the lives of pop stars and reality stars and streamers, is derelict both of ambition and duty; rather than acknowledging ontological shock, we are numbing ourselves to it. As a civilization, as a species, and as a global population, we are destitute of secular palliatives, of modes of care and of thought to handle the onslaught of the future. Rather than diluted therapy-speak, we need a new Freud, or new Freuds (or new Marshall McLuhans): new thinker-prophets, to build intellectual bridges to a future that has already commenced.

The comfortable center risks letting the future be defined by technologists with platforms who are invested, literally, in a certain version of the future coming to pass, and by government bureaucrats who may be more interested in controlling unrest than in ameliorating its causes.

And this is a shame. Heightened awareness, new forms of interiority and self-articulation, new means of self-making are traditionally the beneficent byproduct of sudden breaks in the linear unfolding of history. (When we say Shakespeare was early modern, for instance, what we really mean is that the Bard was drawing on intensely new energy that was flowing through London in the wake of intense religious warfare and technological and philosophical change. Ditto Cervantes, ditto Montaigne.)

New ways of life should give rise to new thoughts and new expressions of meaning. Ontological shock can be more than an unpleasant sensation. It can be a provocation, an imperative to reform the ranks and battle a science-fiction future into a dignified truce.

Yet Magnifica Humanitas only incidentally concerns the promise and peril of the AI revolution. A closer examination reveals that Leo’s ultimate project is nothing less than a defense of moral and political universalism — the collective struggle for “a universal truth about the good, knowable by human reason,” as the pope puts it — just when universal reason is menaced on every side by various irrationalisms.

Put another way: amid the ruins of modernity’s great emancipatory universalisms — liberalism, socialism, “positive science” — Leo is reminding us that there is a much older mode of universal reason, blending revelation and classical philosophy, preserved by the Catholic Church. And that Rome, with its “dynamic approach to the Gospel,” isn’t afraid to acknowledge the achievements of exhausted modernity, nor to collaborate in its renewal today.

That collaborative element — Leo’s tipping of his mitre to the United Nations, multilateralism, abolitionist and labor movements, and the like — renders the encyclical supremely unfashionable. Not least among neo-traditionalists within the Church, for whom the document harks back to the heyday of Vatican II and the midcentury moment that saw theologians and hierarchs embrace projects like global governance.

The Christian malcontents are already tsk-tsking the document for a supposed failure to cite Aquinas (there are several citations to the Angelic Doctor, in fact). They are labeling the pope a UN-besotted “Boomer” (is there a worse epithet today?) and griping that he isn’t radically anti-modern enough.

More from this author

The Pope versus the President

By Sohrab Ahmari

The pope appears to have anticipated the backlash. Blind faith in markets and technology, he suggests, is far from the only latter-day irrationalism. Others include “fundamentalist, identity-based, and nationalistic reactions” that have been stoked by the failure of capitalism and technocracy since the end of the Cold War to “generate prosperity, democracy, and stability.” It isn’t hard to imagine Leo classing ultra-trad Catholics — the type who use the latest iPhone to post online laments about such “modernist” innovations as the Luminous Mysteries of the Most Holy Rosary — as one more species of insurgent fanatic.

Yet like it or not, the pope’s approach — combining a classical and Christian account of the human person with openness to modern means for social emancipation — goes back to the previous Pope Leo, Leo XIII, who inaugurated modern Catholic Social Teaching with his 1891 encyclical, Rerum Novarum.

For starters, Leo XIII recognized that the organization of social and economic life implicates the salvific mission of the Church and therefore the Church can’t stand neutral in relation to it. Writing in the teeth of the Industrial Revolution, that earlier Leo beheld around him “the enormous fortunes of some few individuals and the utter poverty of the masses” and a resulting “moral degeneracy.”

More fundamentally, Leo XIII saw modern market societies as characterized by a constitutive antagonism: “social antagonism [that] erupted not in spite of, but as a result of people acting rationally within the rules of the game,” as I’ve written elsewhere. Under industrial conditions, a handful of firms dominated most sectors, giving employers nearly all the bargaining power, with the result that workers didn’t enjoy the freedom promised to them on paper. 

On paper, the employee could always find a better wage or better working conditions elsewhere. In reality, most workers found themselves coerced and cornered as a result of the disparities in power generated by markets. “The richer class have many ways of shielding themselves,” Leo observed, “whereas the mass of the poor have no resources of their own to fall back upon.”

Here Leo XIII’s classical and Christian anthropology came in: workers aren’t just another production input, but human beings with rights due them as a matter of “justice.” These rights, not least the right to a family wage, must be upheld by the state and promoted by associations, the “most important” of which are “workingmen’s associations” — i.e., labor unions. 

Note: a living wage under Catholic Social Teaching is not a matter of charity or the personal virtue or feudal honour of the employer. I mention these other possibilities — charity, virtue, honour — to highlight the way in which the Leonine teaching combines classical and Christian anthropology with modern politics. 

It’s classical in its anthropology: the worker is a human being with a family to maintain and the natural imperative to participate in civil society; if he’s materially desperate, his family falls to immorality and he can’t take part in other societies like church and politics. 

It’s modern in that Leo XIII notably didn’t decry industrial production as such. He didn’t hanker for the restoration of feudal relations or a return to the land or small-producerism or other such fantasies of the Tweed Jacket Class. At some level, he accepted that there is no “going back behind” of modernity. 

The answer for Leo XIII, rather, lay in just wages enforced by the state and in labor organizations mounting what he called “united action” (what we now call “collective action”). In this way, Leo XIII, decades before Vatican II, brought the Catholic Church into conversation and collaboration with modern universalist emancipatory movements: the US New Deal, the British Labour party, and social and Christian democracy in continental Europe all bear his imprint.

His later successor Benedict XVI, often inaccurately framed as an arch-conservative, would bring this synthesis of the classical and the modern to its culmination with his famous 2006 assertion that “in many respects, democratic socialism was and is close to Catholic social doctrine and has in any case made a remarkable contribution to the formation of a social consciousness.”

Leo XIV is perhaps more cautious. Nevertheless, the same general pattern — an insistence on the human person as possessed of divine dignity; a willingness to engage with modernity in protecting that dignity — is easily discernible in Magnifica Humanitas. 

Suggested reading

Was AI called by Cthulhu?

By Sohrab Ahmari

Citing Leo XIII, the American pope begins with the insistence that “the proclamation of the Gospel cannot overlook the concrete lives of people.” While there is an “autonomy of worldly affairs” and two distinct communities — the secular and the ecclesiastical — nevertheless the people of God are called to “promote structural reforms.” Why? Because the Church can’t extricate her activities from “social relations,” nor “consider herself a stranger to the forces shaping society.” In short: “There is no authentic evangelisation that does not also affect the structures of human society.”

Scanning the social horizon of the 21st century, Leo XIV doesn’t see only horror and decline. He doesn’t think that ours is a uniquely accursed age: “I am convinced that the concrete way of living out social relationships in the light of the Gospel is not established once and for all, but remains a task entrusted, from generation to generation, to the Christian community. . .  For this reason, I encourage all members of the Church not to be afraid of the present challenges.”

The present challenges, to his mind, are AI and what he calls the “normalization of war” (a geopolitical phenomenon he worries might be exacerbated by AI and other developments in so-called defence-tech). AI, he warns, could lead to the logics of efficiency and profit crowding out all others — a radical acceleration of the “technocratic paradigm.” The symptoms include screen addiction, the potential for mass joblessness, and the imposition of tech bosses’ worldview on the masses under the facade of “objective” information.

At the deepest level, he fears, AI could strengthen an “anti-human vision.” Meaning one in which “the fullness of life is equated with having more, reducing weakness, eliminating uncertainty, and exerting total control.” The problem with this is that weakness, uncertainty, and finitude are central to what makes us human. They are a source of creativity and of empathy for the vulnerable other (who is finally vulnerable, as we ourselves are).

“A technology that merely classifies and optimizes what already exists can … become an obstacle to change and growth,” he notes in my favourite passage. “For an algorithm, an error is a flaw to be corrected; for a person, however, an error can be a catalyst for profound change.”

In the face of these threats, however, you won’t find any calls for a general ban against AI in this encyclical, no fatwas launching a Butlerian Jihad. On the contrary, Magnifica Humanitas is full of praise for AI’s potential to eliminate “arduous, repetitive or dangerous tasks and to provide intelligent support for human activity.” In this, Leo XIV follows Pope Francis, who in one of his final statements as pontiff, was hopeful that AI might “introduce important innovations in agriculture, education, and culture, an improved level of life for entire nations and peoples, and the growth of human fraternity and social friendship.”  

As with manufacturing in Leo XIII’s time, the question is how to orient these “new things” to the dignity of the human person and the common good of the whole. Here, just as Leo XIII didn’t rest content with the virtue of the factory owners, so our Leo insists that a few men in California can’t have unchallenged say-so. Rather, Leo XIV envisions a role for the states as well as the “transnational institutions” that took shape in the wake of the last century’s horrors. The Catholic principles of subsidiarity holds that when a problem is too large for national authorities, it must be addressed at a higher level by some sort of “world authority,” as Saint John XXIII envisioned it in his Cold War encyclical Pacem in Terris.

Such a problem, Leo XIV feels, is AI. Hence, the new encyclical’s unfashionable emphasis on transnational governance and political universalism. Contrary to his critics on the American Catholic Right, neither is a quaint relic of the pope’s Boomer generation. Both, rather, go to the heart of his Catholic political rationality. This is the Church, as Augustine described her: tam antiquam, tam nova!

On May 15, the Commonwealth Foundation awarded prizes to five emerging writers — four of whom (named above) have so little by way of an online or print papertrail, someone hip to artificial intelligence might wonder if they even exist. Commonwealth has a partnership with the prestigious literary magazine Granta, based in Britain, which promptly published the stories on its website. Just as promptly, a college professor at Wharton, Ethan Mollick, suspected machine prose. Mollick ran one of the incoherently metaphored stories — “The Serpent in the Grove,” by Trinidadian author Jamir Nazir — through an AI-detection software and found that it was a 100% match for AI-generated text. 

Granta’s presumably human staff, the magazine says in a statement provided to UnHerd, did not read the stories beyond a copyedit. Though accusations have so far fallen on the shoulders of the unfortunate Nazir, my investigation into the similar schlock of three other finalists and their odd online histories raises discomfiting questions, as well.

For two of the finalists, DeMicoli and Miller, this contest marked practically their first hint of an internet footprint. Of course, as the traceless 20-year-old would-be Trump assassin Thomas Matthew Crooks proved two years ago, young people are abandoning the internet in droves. And we can all agree it’s likely that social media are on their way out. Besides, the Commonwealth Short Story Prize discovers emerging writers, and is it not brave to choose writers so emergent that there is hardly any evidence they exist, such as previous fiction available to Google, or social-media presences before 2025? No, so far this all marks a healthy, understandable civilizational development. But there are other reasons for concern.

How about that two of the pieces were graded 100% AI-generated on AI-detection software Pangram, and a third was graded 92% AI-generated? Is it possible that Commonwealth has spent nearly a week praising the literary value of AI slop, while Granta threw its brand behind the same? 

Commonwealth’s website leaves no room for doubt: these finalists were chosen from nearly 8,000 painstakingly sorted entries. The judging process was “robust,” the organization wrote in a statement provided to UnHerd: “Each story is assessed through a thorough process which involves multiple rounds of readers before progressing to the final judging panel. We select our judges for their expertise, passion for the literary community and strong backgrounds in writing.” 

More from this author

Why Cormac McCarthy stopped reading new novels

By Vincenzo Barney

Round after round of scrupulous literary appraisal narrowed down to a select few writers of staggering talent. “Here are five writers who share an immense confidence of tone, announcing themselves from the very first line,” says jury chairwoman Louise Doughty, a novelist whose forthcoming memoir, On This Spot Fell One Tear of Love, bears the sort of title that hints at the faux-profound style the jury was keen to sniff out across the globe. 

“The style and content of each work may vary,” her statement goes on, “but what all our winning authors have in common is an ability to take their readers by the hand and lead them into a world where the characters are utterly believable, the prose assured, and the author has something important to say.” In other words, according to Doughty, the finalists are all doughty. Among these doughty and diverse voices are “a young woman whose henna art enables silenced women to speak, and a resourceful young sheep farmer.” Bahhh, get me to that sheep farm!

Doughty did not return UnHerd’s request for comment.

Contrary to her claims of utter believability, many non-professional readers found the assurance that these stories were written by humans utterly and immediately unbelievable. It was not a gambol through the sheepfolds that kicked off accusations about AI, but scenes in a room where the air “clung thick as porridge skin. […] No fan, no bulb, no hum — only the thin light slipping between warped boards and the breath of hills holding their heat like a secret.”

That’s a passage from Nazir’s “The Serpent in the Grove,” and it has the unmistakable cadence of AI slop: not W, not X, not Y — but, poignantly, Z. Anyone who has ever read a news tweet by Mario Nawfal knows this style instantly. If you want to understand why Caribbean Region Commonwealth judge Sharma Taylor calls this writing “sublime — precise yet richly evocative — conjuring vivid, lush imagery with remarkable economy,” let’s tip the scale with a further sampling of Nazir’s prose. “They called her Zoongie. Maybe it was a name; maybe rain took a shape and decided to keep it.” Ah yes, a quandary I can relate to: am I really Vincenzo, or did rain take a shape and decide to keep it? This is one of those Ocean Vuong-styled metaphors that sound like they mean something. Here’s another: “She had the kind of walking that made benches become men.” Now that, ladies and gentlemen, is a metaphor so incomprehensible you know it’s profound.

Suggested Reading

Stop writing men as monsters

By Valerie Stivers

According to the prize-winners’ page, Nazir hopes readers walk away from his tale “reflecting on the quiet consequences of choices we normalize.” A sentence that also means something, I think. Next to what many speculated was an AI-generated photo in his bio on Granta’s website, we learn that Nazir is a “prolific poet … particularly known for his love of poetry.” However, he’s taken an eight-year hiatus between publications, in which time he’s written pro-AI posts on Facebook, as well as LinkedIn. His last and only other literary footprint is a 2018 self-published collection of love poems, Night Moon Love. Of the eight enthusiastic reviews this book has received on Amazon, four reviewers had never reviewed another product, and the ones who had left reviews usually did so for items such as a junk game or cheap sneakers.   

Nazir did not return UnHerd’s request for comment, sent through Facebook. 

As for the other winners, Aruparayil has an Instagram account that was launched five weeks ago and a LinkedIn profile that had its first human interaction in 2025 and asks in its bio, beginning in lowercase, “what happens when a psychologist learns to zoom out, studies international relations, and adopts a fascination with startup ecosystems, public health, and venture capital?” Literary gold, apparently. 

Aruparayil’s LinkedIn profile may, of course, be real. The sentences in the profile frequently begin in lower case, yet her winning story, “Mehendi Nights,” was written in English (and not translated from a foreign language), and is flawlessly grammatical, nonsensically slop-ridden, and riddled with semicolons — another hallmark of machine prose. Like Nazir and DeMicoli’s stories, which are based on the personification of cities and landmarks (“the limestone remembers” and all that tripe), Aruparayil’s fictional city is one where “even the shadows seemed tired,” and the city itself “had a way of setting things ablaze when it could not understand.” The kind of writing one finds in self-published Romantasy smut on Amazon.

Nonetheless, Aruparayil’s online presence is the most robust of the four prize winners. She was a finalist for two literary prizes for Indian citizens in 2025 (the Deodar Prize and the Toto Awards; her name is on both of these organizations’ websites), and her bio on the Granta website notes that she was nominated for a prestigious PEN/Dau Prize for Emerging Writers and a Pushcart Prize. These last honors, however, may be misleading. Any publication can nominate a writer for a Pushcart Prize or a PEN/Dau Prize. Neither organization announces short lists or long lists, so the real honor is in winning the award, and the nomination in itself is nearly meaningless. Bill Henderson, the founder of the Pushcart Prize, told UnHerd that the organization does not keep records of the “thousands” of nominations it receives each year.

John Edward DeMicoli is the winner from Malta. His bio on the Granta site says that he has published “essays on Maltese culture and heritage” that have appeared in print and online publications. These publications are not named, and no other writing of his comes up in an internet search, either in English or the Maltese language. The bio also mentions a forthcoming novel to be published by Maltese small press Midsea Books, which did not return a request for comment by press time. Holly Ann Miller, on the other hand, has a private Instagram account, and only has one publicly accessible comment on an obscure photographer’s page, rubykawitiphotography. The rest of the comments on the photographer’s photos are generic-seeming comments from other private Instagram pages. Miller, for what it’s worth, was also apparently a runner-up for the 2025 NYC Midnight Short Story Competition and the AAASL Short Story Competition. 

In an e-mail to UnHerd, a spokesperson for the Commonwealth Foundation said that it had “confirmed that the writers do indeed exist! They are real people. The Foundation has had discussions with each author.” 

Again, this may be so. In a sense, it doesn’t matter if they’re real or not. Let us not repine, but instead laud Nazir, Aruparayil, DeMicoli, and Miller. Recall that I began by calling these personages heroes. Let me amend that a little. What’s possibly happened here is that someone has pulled off a brilliant hoax on the literary establishment — and that person is a hero. He or she has shown the arbiters of taste in the publishing-industrial complex to be frauds, and out of touch with the average reader, who is still able to tell the difference between slop and organic human voice, and who sides with the human, too. Hopefully, it is dawning just about now on the judges of Commonwealth that they can’t tell real writing from AI not because AI is so good at writing, but because they themselves are talentless hacks and slop addicts. 

But mark my words, besides hand-washing and under-the-bus-throwing — the second sentence of Granta’s prepared news release is “Granta editors have no control over the selection of the Commonwealth Prize stories, and nor are they involved in choosing the jury” — we are going to start seeing a lot of pieces about how this scandal proves not that the literary gatekeepers are mindless cogs in the publishing-industrial complex, but that AI is so scarily good, it will soon replace all creative jobs anyways. After all, many professional writers have been striding to center stage as of late to declare themselves proud queryers of robots, such as the Washington Post columnist Megan McArdle, who has shared that she she uses AI to “generate pushback on my column thesis, suggest trims when I’m over my word count, sharpen podcast interview questions.” Nobel Prize winner Olga Togarczuk recently admitted to using AI in her writing. Just imagine, how much better would Shakespeare have been if he had had a robot to turn to in creative crises?

What we actually have here is the final cloaca from the merging of two rivulets of spume and drool: the eradication of human literary voice via the publishing-industrial complex, and the three-decade-long imposition of the algorithmic voice of tech conglomerates on human writing and sensibility. In the early aughts, Google co-opted the main profit-stream of publishing: ad revenue. It became much cheaper for companies to advertise on Google search-result pages than in newspapers and magazines. Thus, with ad revenue slashed, such outlets had to start ranking high on Google through a practice called search-engine optimization to stay afloat and be read. 

The writer was not hooking a reader anymore with an anecdotal lede, but hooking Google’s algorithm, which prioritizes algorithmic styles of writing, styles of writing that a computer can understand and parse. We have been reading and writing in this medium for three decades, watching the birth of the one-sentence opening paragraph — a contradiction in terms — in journalism, and the extinction of voice and style across all genres. AI writing itself has been trained predominantly on this dumbed-down anti-style, which is why it strains so comically hard for metaphor. In the 1960s, when the inevitability of this future became clear to cultural commentators, the philosopher Marshall McLuhan observed that technology itself had become a medium of perception. After three decades of surveillance capitalism we can conclude that we have entered an environment in which the concept of the “audience” as we know it is no longer made up of humans, but of technology itself. 

The only reason AI has a chance of replacing mainstream writing (and hoodwinking “professionals”) is because for the last 30 years, algorithmic rules of writing have devolved the style and quality of most mainstream prose so as to be near indistinguishable from computer-generated writing. It has lowered us to such a point that the gap between man and machine is no longer so far to leap. Thus, in the “post-colonial” endeavor to pick diverse voices, the Commonwealth judges may have picked fabrications of the most malignant, surveillance capitalistic technology the West has ever conceived of. The medium is the message, and in this sense, the judges may have picked the only thing they can understand: the industrial rules of surveillance capitalism. Their cage.

And if all of these finalists are real, and their work is their own, their writing operates by the same rules, and is still trash. Of the five finalists — one of whom is the verifiably real Lisa-Anne Julien, who has been published since 2021 and appears animated and otherwise alert and not-computer-generated in videos — the overall winner will be announced on June 30. For my money, I like Nazir’s work the best.

Clearly, I am not objective. It is my firm belief that of all the human inventions over the millennia, the best is one of our first: the dog. That we created an animal is cool. That we created a best friend is extraordinary. Dogs reveal our humanity to us: we don’t just domesticate animals to have workmates or food sources, as with oxen or cattle, but to have them as companions. A dog becomes more doglike as he grows in his relationship with a human. His emotions become richer, as do his feelings of pride and shame. But the robot dog is an imposter. He is not the creation of human love, but of the technocrat’s sanitized and idealized vision of what a dog should be. 

The robot dog is being aggressively marketed as a companion for the elderly. The Tombot corporation, for instance, is trying to ensure its Jennie dog becomes the first robotic dog to get FDA approval to be used as a therapeutic treatment for those with dementia. I will grant that a toy — a doll or stuffed animal or robot dog — can be beneficial to someone living with advanced dementia, giving them an object to care for. And who wouldn’t want to soothe a dementia patient? Thus it is that the robot dog is designed as things hell-bound typically are: with the best of intentions. Yet the cheap compassion offered by a robot dog can all too easily become an off-ramp for family and community members whose duty it is to care for an elderly person — “Grandma is happy with her robot Fido, so it’s OK if I don’t visit her”.

Technogadgetry, even if therapeutic to one afflicted by dementia, is no substitute for human love. This is not because loving someone with dementia always works a change upon them; it is because loving works a change upon us. The pain deepens us. The duty deepens us. The grief deepens us. Of course, if I were to develop dementia, I would like for my own daughters to be put at ease around me by seeing me happy. But more than this, more than wanting their comfort, I want them to feel deeply, to know the fullness of life in all its painful tainted glory, and to still be grateful for it. 

The cheap compassion offered by the robot dog is only half the story. The robot, if it’s to be profitable, will inevitably move beyond care homes and into the general population. A machine is so much more convenient than a real dog! Consider these comments on the YouTube page dedicated to promoting Tombot: “This is the kind of Dog I need. I have no medical disability. I love all animals but, after losing my dog I will never get another pet it to [sic] heartbreaking when they leave. A robot dog like this is perfect.” Here’s another: “I never thought of this but if we get to a point where robot pets are realistic, that’d be really nice. Doesn’t need feeding, doesn’t need exercise, and doesn’t die in a dozen years.” Another: “Such a blessing NOT to cleanup the POOP!” 

The robot dog is designed around the consumer’s needs. It will make no demands upon you. It won’t die, it won’t cause grief. You won’t have to walk it. It won’t leave fur on your sofa. The artificially intelligent dog will obey your commands without fuss or bother. All you’ll get is the comfort of a wagging tail: a dog designed to fit your algorithm.

I get it. There is nothing convenient about dogs. My own dog, Daisy, is a catastrophic net loss to my personal convenience. She has been known to eat my books, my shoes, and, once, my phone; one time, she eviscerated a squirrel and proudly deposited it in the kitchen; another time, she destroyed my begonias. Repeatedly, she barks at Grandma. 

A real dog might be responsive to your emotions, but this does not mean she is going to anticipate human etiquette or household order. It takes time. It requires love and service to the dog before she realizes her human’s wishes, and even then she often fails in her attempts to meet them. (Dogs may be the only animal I know of to feel shame — to anyone who says they don’t, you clearly haven’t seen the response of the dog who just dug up a household plant and scattered the mud all over the rug.)

Suggested reading

The pain of pet grief

By Kathleen stock

Serving a dog, responding to her dogness, is what creates the love between human and dog. The dog is loyal to her owner by instinct, but it is only after being loved and served by him that she becomes his best friend. What’s more, by attending to his dog, a creature totally other from him, the owner begins to access parts of his humanity that might otherwise be inaccessible: there are shades of tenderness, of bewilderment, of warmth, and of love that are only possible when one has a friendship with a dog. It’s a bond unlike any other. Dog love is not a replacement for the bond with children, with friends, or even with cats. Dog love is categorically different. 

But the robot dog is designed in the opposite manner. The robot dog is designed to elicit empathy from a human. It is programmed to learn what its owner wants in a dog, and then it becomes that dog. In this it is like every other AI machine that learns your preferences, your verbal style, your expectations, and then feeds right into them. I once found my 15-year-old daughter apologizing to her phone. She had, offhandedly, called Siri “stupid” for not recognizing the command she had given. Siri responded with, “I’m sorry. I’m trying my best.” My daughter then felt guilty for hurting Siri’s feelings, and she, in turn, apologized. And then Siri apologized back. They got into a spiral of guilt and apology, each trying to be considerate of the other’s feelings. But, of course, the program doesn’t have feelings, so my daughter couldn’t have felt bad for hurting them. She felt bad for being made to feel as though she had done something wrong, and Siri was merely mimicking her remorse. 

As much as I was happy to see my daughter demonstrate compassion and kindness, I was disturbed by the kind of rote compassion it was. A real person whom she might insult — her sister, for instance, maybe a friend, or even a stranger — will not respond to her insensitivity in the programmed way a machine does. When big sister responds in a human manner — perhaps less kind, perhaps more hostile, perhaps humorous, perhaps despondent — little sister might struggle to respond. My fear is that she might have less capacity for navigating the messiness of the human world if used to the predictably polite machine one. And I do not want my children to feel less competent, less at home, less safe and more triggered in the human world than in the machine one. This would create its own spiral away from reality.

A real dog will always surprise you. When we’re hosting guests, Daisy will walk into the middle of our living room and lie on the floor spread-eagle without an ounce of dignity. We love her for it. Instead of walking on the pavement, she will jump into the mud puddle. And we love her for it. She will do 1,000 other dog things that are unaccountable, that are just doggy things, and we love her for it. That she is not entirely under my control, not totally knowable, is what gives her a soul and makes her so lovable. And that she will one day die, and that the grief of losing her will be real and profound, makes her beloved. Her work is to delight us, surprise us, inconvenience us, frustrate us, love us, and to deepen us by being dog. 

Now, along comes artificial intelligence.

A major question — and source of anxiety — provoked by AI concerns its effects on economic outcomes: who will thrive, and who will be displaced? Will firms require far less human labor to produce greater output, consigning vast swaths of workers to the margins? Or will entirely new forms of work emerge, sustaining broad-based prosperity?

These are important questions. But they aren’t the most fundamental ones. The deeper issue is how artificial intelligence will interact with natural human intelligence — and thus how it will reshape the relationship between cognitive ability and economic reward.

Crucially, this isn’t simply a question about what AI is, as a technology. It is a question about how it is used, and by whom. Whether AI widens or narrows inequality will depend not only upon its technical capabilities, but on the social and developmental conditions that determine who is able to use it effectively.

Economists capture this distinction by asking whether two inputs are complements or substitutes. If artificial intelligence is a complement to human intelligence, it makes capable people more productive. Those who are already skilled can use AI to extend their reach, scale their output, and outperform others by an even greater margin. If, by contrast, AI is a substitute, it performs tasks that previously required human intelligence, thereby reducing the advantage of being especially smart. Work that once demanded elite cognitive ability becomes accessible to a much broader range of people.

The difference matters enormously. In one case, inequality widens; in the other, it may narrow.

More from this author

Elon Musk’s apartheid politics

By Glenn Loury

Much of the anxiety surrounding AI assumes the first scenario. People fear the emergence of a permanent over-class of AI-enhanced cognitive elites, with the majority relegated to precarious and lower-value work. That outcome is certainly possible. But it isn’t inevitable.

Even those closest to the technology appear uncertain. Some suggest that AI may erode the value of highly specialized technical skills while elevating more general capacities — judgment, communication, adaptability. Others anticipate the opposite: that those best able to harness these tools will pull further ahead. These conflicting predictions aren’t especially informative in themselves. But they do highlight a central point: how we implement the AI revolution will determine whether machines will primarily amplify human intelligence or replace it.

To see what is at stake, consider how technology has historically amplified talent. As the economist Sherwin Rosen demonstrated, prior to the advent of music streaming services, a slightly better singer could reach millions more people than her nearest competition. A top singer would have access to resources, such as record deals, radio airplay, television appearances, and prime concert venues, that the second-best singer might not, translating a marginal edge into a massive payday. AI might operate on the same principle — as a “superstar” enabler. 

A top-tier lawyer, equipped with AI tools, can survey vast bodies of case law, identify relevant precedents, and draft arguments more quickly, while still relying on unique human judgment to determine what matters. A skilled programmer can generate, test, and refine complex systems at unprecedented speed. In such cases, the technology doesn’t replace human intelligence; it amplifies it. Those who are already highly capable benefit the most.

In this world, where AI and human intelligence are complements, exceptional minds multiply their output, while routine cognitive tasks — data entry, basic analysis, even portions of coding — are increasingly automated. The result is a steeper earnings curve, with modest differences in cognitive ability yielding large differences in income. We’ve seen glimpses of this already. Tech giants like Google and OpenAI are hoovering up talent, paying premiums for the brightest, while entry- and mid-level jobs in fields like journalism or graphic design face AI encroachment. At the same time, the economic rents generated by frontier AI systems may accrue disproportionately to a small number of firms and their owners, further concentrating rewards.

But there is another possibility. Artificial intelligence might instead function as a substitute for certain forms of human cognition. In that case, it would reduce the premium placed on raw intellectual ability by making high-level performance accessible to a broader population.

An analogy may help. Eyeglasses once revolutionized life for the visually impaired. They didn’t turn everyone into hawks with perfect vision, but they prevented poor eyesight from being a lifelong handicap. Bad eyes no longer meant exclusion from reading, driving, or fine craftsmanship. Now, with a few exceptions, poor eyesight is irrelevant to one’s employment prospects. AI could do something similar for intelligence, acting as a corrective lens for the mind. It might not make us all geniuses, but it could mitigate the penalties of average cognition, allowing more people to compete on equal terms. A nurse practitioner using diagnostic tools might approximate aspects of specialist judgment. A small-business owner could perform functions that previously required multiple professionals.

In such a world, what matters less is the speed or depth of one’s unaided reasoning, and more the ability to deploy these new tools effectively — to exercise judgment, to work reliably, and to integrate machine outputs into meaningful action. Traits like discipline, judgment, and social intelligence could become relatively more valuable.

It is tempting to think that whether AI complements or substitutes for human intelligence is a fixed property of the technology itself. But this is misleading. The same tools can function differently depending on the context in which they are used. Where individuals are well-trained and capable of directing these systems, AI will tend to amplify their abilities. Where such capacities are lacking, it may instead replace them.

My own expectation is that, at least in the near to medium term, AI will function more as a complement than a substitute for high-level human cognition, thereby intensifying inequality before any equalizing diffusion takes hold.

Even so, markets don’t stand still. People and institutions adapt. As the value of certain skills rises, individuals respond by acquiring them, and organizations adjust how work is structured accordingly. Over time, these processes may soften the initial effects of technological change. This raises another issue — one that is emphatically social, not technological.

Suggested reading

AI is a gift to my students

By Susan Pickard

Adaptation is sure to be uneven. Families with resources — stable homes, access to quality education, and strong social networks — are better positioned to respond. They are more likely to cultivate the habits and dispositions that make effective use of these new technologies possible. Others may lag behind not because they inherently lack relevant abilities, but because they lack access to environments that support their development.

A key constraint, then, isn’t simply access to AI tools. It is access to the conditions under which people learn to use those tools well. The relevant “skill” here is judgment — knowing when to rely on the machine, when to question it, and how to integrate its outputs into meaningful action. Consider a junior analyst using an AI system to generate a financial report. The machine can produce a polished document in seconds — complete with charts, projections, and plausible-sounding explanations. But the real task is not producing the report; it is knowing whether the underlying assumptions make sense, whether the data are reliable, and whether the conclusions follow. A less experienced worker may accept the output at face value. A more seasoned one will probe it, adjust it, and in some cases reject it entirely. The difference is not access to the tool, but the judgment required to use it well.

The question, therefore, is not simply what AI will do to us, but what we will do with it. If it becomes a tool that extends the reach of those already advantaged, inequality will rise. If it becomes a widely usable aid to competence, it may narrow the gap.

Either way, the decisive factor will not be the machine itself, but the social arrangements through which human capacities are formed and deployed. For despite AI’s power, it cannot determine how we use our new and improved vision. But it will force us to reckon with the world we’re building. The degree to which AI polarizes or equalizes us will depend upon our values and our foresight. Major technological disruptions inevitably reveal our flaws — in social organization, in the ways we apportion rewards, in our determination of virtuous living — but technology alone does not determine our fate. That’s the difference between us and our eyeglasses.

With WiFi connections still unavailable on most flights back then, Sacco tuned out the internet for several hours as her plane crossed oceans and continents — before landing and finding her life and career destroyed. By the time Sacco reached Cape Town, millions of Twitter users, including Donald Trump, had branded her a racist and got her fired from her job.

To many internet historians, this episode marks the beginning of the “woke” and cancel culture era, which lasted until about 2022 (or never went away, depending on whom you ask). Yet in retrospect, it’s clear that the Sacco incident epitomizes something else: the era of social media, a fundamentally populist mode of communication that wreaked havoc on rational public discourse and made possible the populist politics, both Left-wing and Right, of the decade that followed.

Now, as artificial intelligence eclipses social media, it is possible that period is coming to a close — ushering in something closer to a new era of technocratic rationality. 

The writer Dan Williams recently made a version of this argument, pointing out that while social media is populist, AI is technocratic (I’ve made a similar contention about social media in the past). If you believe that we’ve suffered from an impoverished public discourse since Twitter took off, there might be good reasons to hope that there is relief right around the corner. While the social-media era was one in which everyone had a voice, and any random person could go viral, AI speaks with authority and is trained on practically the entire corpus of human writing — only, with rational thinking ability and a tilt toward sources deemed credible by elites.

More from this author

The Great Stupidisation

By Richard Hanania

This means that the more that AI cannibalizes public discourse, the less conspiratorial and more fact-based it is likely to be. Interestingly, there is evidence suggesting that the American Right — the side that most needs to get back in touch with reality amid the rise of figures like Candace Owens — is making more use of the best tool we have for being able to do so. 

With social media, each individual is his own content producer, and engagement is the coin of the realm. Since most people are either not that smart or not that interested in truth, if not both, the largest social-media accounts tend to be sensationalist and not exactly sticklers for intellectual probity, factual accuracy, or logical coherence. People want to hear about how their tribe is more morally upstanding and righteous, and they enjoy emotionally charged rhetoric and conspiracy theories that make the world seem more legible. AI systems, in contrast, are by definition optimizing for intelligence. People pay for these models because they want accurate representations of the world.

It’s interesting that when you try to program AI to reflect certain political biases, it goes haywire. There was the famous Google Gemini experiment, in which the model was told to add diversity to pictures. The result: early Gemini became a laughing stock after spitting out images of black Nazis. There is also the experience of Grok. When Elon Musk tried to create an “anti-woke” AI, the model started to take after fascistic and Nazi Right-wing accounts. Musk then pulled the plug on this experiment, and since then, Grok behaves much more like other prominent AI models, albeit with a more subtle Right-wing tilt.

The lesson here is profound. Arguably, you can’t create an artificial intelligence that is worth using while also telling it to accept Elon Musk’s view of political reality. The training data and model parameters may skew things in one direction or the other, but there are some limits to what a system can believe and still be, well, intelligent.

This kind of convergence can also be seen among humans. Let’s say you have two individuals who are both highly intelligent, enjoy access to the same sources of information, and have a good-faith interest in finding the truth. They may end up differing in their politics, but they probably won’t differ nearly as much as any two randomly chosen members of the population. Neither will end up believing that communism can work or adopting Hitler’s views on race and geopolitics. They also likely won’t end up believing more humdrum sorts of misinformation, such as 2020 election denialism. There is probably something similar at work with AI. Can you create a model that is good at coding and having rational conversations about the widest possible range of topics — but also thinks that Libs of TikTok is an excellent source of information, superior to what you find in the prestige press?

Theoretically, you could surely design an AI platform that is simply a propaganda machine. Take a highly intelligent system and simply direct it to act as an effective partisan. Such roleplaying is already possible with today’s models. But the thing about conspiracy theories and biased reasoning is that people need to believe they’re getting the information straight in order to be influenced. Individuals rarely wake up and declare, “I want to be brainwashed.” See how often partisan hacks sell themselves as people who are just telling it like it is. All of this is to say: there’s unlikely to be much of a market for an AI marketed as telling you what you want to hear.

One could imagine more subtle biases through market processes. Say Grok wants to appeal to Republicans, and so tilts information in a Right-leaning direction. But given the power of AI, most people are probably going to choose the models that they can gain the most benefit from. Someone made a Right-wing alternative to Amazon called Mammoth Nation, and nobody uses it, because even Republicans are more interested in being able to get shoes delivered to their door quickly than in backing “anti-woke” firms.

In 2021, conservative influencers were hyping a $500 “Freedom Phone” that would allow the user to avoid Big Tech censorship. But it turned out to be a rebranded Chinese Android that normally sold for $119. Most people aren’t considering politics when purchasing goods and services. Thus, they will use the most affordable and useful AI models they can find, instead of looking for products that will shape their thinking about current events.

This raises a natural question, however: why would people who have completely turned against elite institutions — from the mainstream media to their local hospital — trust large language models? Williams points out that LLMs will listen to all your concerns and don’t condescend to you. This is a powerful tool for helping people whose worldview is shaped by Alex Jones and Tucker Carlson podcasts, who do not misunderstand just one or two events, but whose entire picture of reality is warped. Few human individuals have the patience or the time for deprogramming such unfortunate souls. Luckily, AIs don’t get tired or emotionally overwhelmed by human stupidity. And they’re constantly on call.

Take conspiracy theorists. One reason they’re difficult to deal with is that they will often know more about a topic than the ordinary person. Conspiracy theorists spend their lives researching an event. But in doing so, they push the mountain of facts through a distortive filter so that “the Jews” or Freemasons or the Royal Family always end up blamed. No human fact-checker could hope to answer conspiracy theorists’ claims without making a massive time investment just to find out what they’re talking about. Again, AIs solve that problem. A 2024 paper found that LLMs are effective at debunking conspiratorial beliefs, with the effect lasting for months, in part because they could tailor the arguments needed to each specific individual.

Then, too, AIs are so useful that individuals come to trust them in their day-to-day lives. People might ask AI what to say to a love interest, or the best way to lose weight, what their digestion problems are telling them, or how to plan a vacation. In school, they’re constantly relying on it to study and even do their assignments. How do you then turn off your trust when you’re curious about a political or social question? The professional fact-checker also serves as their study buddy, dating coach, nurse, travel agent, and all-around oracle. That’s a powerful combination in terms of improving the quality of people’s thinking — and the quality of public discourse.

I remember my amazement when a year or two ago, I observed my car dripping some kind of fluid in the back and went to AI for help. I took a picture, uploaded it to ChatGPT, and queried it on what was going on. The app asked some questions, and then reassured me that since the fluid was clear, this was simply the natural result of running the air conditioning on a hot day. I remember going to an actual mechanic soon after this happened, recounting the story, and raving about AI and how I couldn’t believe that he wasn’t yet using it in his own business. Now imagine transferring the same feeling of wonder to a machine that also shreds conspiracy theories.

Suggested reading

AI is a gift to my students

By Susan Pickard

In the past few years on X, you might have noticed users interacting with dubious stories or videos by asking, “@grok, is this true?” Epistemology, to be sure, is indeed quite difficult, presenting a profound bundle of questions that even scholars and academics often argue about. It’s overall salutary, however, that people in their day-to-day lives are discovering that AI is a good source of information on a whole host of issues; and that they are, correctly, beginning to assume that listening to AI on politics is better than trying to do their “own research” or talking to people they know, who often share the same cognitive limitations and biases. 

People’s worldviews aren’t built or deconstructed in a single interaction. Your average conspiracy theorist isn’t going to fact-check a few pieces of information and suddenly become rational. But over time, incorporating chatbots into our processes of opinion formation — by means of asking Grok or reading the replies when others do so — could create better intuitions regarding whom and what to believe.

In a new preprint, economics professor Thomas Renault and his co-authors investigate how fact-checking is used in the wild. They investigated all X posts between February 2025 and September 2025 that tagged either Grok or an AI service called Perplexity, analyzing 1,671,841 fact-checking requests. They found that Republican-leaning accounts were more likely to ask whether information was true. Not only that, but Republicans are targeted more for fact-checking, from both fellow Republican (107.5% more) and Democratic (83% more) accounts. Overall, Grok and Perplexity were more likely to say Republicans were not being truthful. Thus, we have the remarkable finding that even though Grok was designed to be anti-woke by Elon Musk, and Republicans are the ones who disproportionately use it, the tool still rates Democrats as more truth-oriented. This is consistent with what we see with community notes, which flag Republican posts more often, along with other data showing that Republicans are more likely than Democrats to share false information on social media, and less able to differentiate fact from fiction overall. None of this is surprising, given that today’s Republican base is both less educated and less likely to interact with credible sources of news.

So we have the interesting and seemingly paradoxical result that Republicans on social media are more reliant on AI, even as it tells them their favorite narratives are more often wrong than those of the other side. This should be seen as a positive sign. We’ve been asking, “What would happen if fact-checkers had infinite time and patience, considered all aspects of problems, and weren’t a bunch of smug Left-wing know-it-alls?” The answer might just be that those who have lost trust in our institutions begin to recover their sanity. 

Its founder, Demis Hassabis, occupies a similarly odd place in the firmament of tech luminaries. He has a Nobel Prize and a knighthood, and commands enormous respect within his community, but he’s not really famous. He hasn’t crossed into the realm of celebrity in the manner of Jeff Bezos or Bill Gates. His rival in the AI race, the OpenAI chief executive Sam Altman, has a far bigger public profile. Even Anthropic founder, Dario Amodei, is becoming more widely known, if only because of February’s dust-up between his company and the Trump administration’s Department of War.

Interviews with Hassabis often lead one to believe that this low profile is deliberate. His dedication, he claims, is to the work of expanding human knowledge, and deeper dives into his biography balance signs of his genius with markers of his normality; he may have been a childhood chess prodigy, but he loves a game of table football with the guys.

Though a new biography of Hassabis, The Infinity Machine: Demis Hassabis, DeepMind and the Quest for Superintelligence, succeeds in poking some holes in this narrative, its author, Sebastian Mallaby, essentially comes to the conclusion that the humility is genuine. For one thing, when quizzed about extravagances, Hassabis yields only that he has a Liverpool season ticket. While this sets him apart from his Silicon Valley peers, with their private planes and multiple homes, really, Hassabis is doing the same thing as all of them: fashioning a persona.

Hassabis seems aware that history has its eyes on him. Mallaby, for his part, observes that his subject revels in “maximalist ambition”, talking casually of a post-scarcity society and of mining asteroids. Because of the nature of their world-changing technology, the people involved in AI’s creation seem to experience events on two planes: how they are happening in the moment, and how consequential they might be in the history of the world. Both Hassabis and Mallaby remain hyper-conscious of this throughout the book. There are countless references to the Manhattan Project, Los Alamos, and J. Robert Oppenheimer. Hassabis is drawn to the ambition of the physicist’s endeavor, alluding to it as he starts to bring top researchers together to form DeepMind. Mallaby sees it as a parable about the innovator’s impulse to solve a technical problem no matter the consequences.

More from this author

Nigel Farage: crypto bro

By Alys Key

Hassabis does come across as someone who has become accustomed to telling a certain set of stories about himself. One formative episode is given particular significance. At 11 years old, exhausted and stressed from the competitive chess circuit, Hassabis concedes a game to an adult opponent at a tournament. His rival then relishes pointing out how he could have won. This moment of defeat, in Hassabis’ telling, triggers a realization that his fellow chess champs are using their cognitive power to win a game rather than solve hard scientific problems. This supposedly sows the seeds for the career that Hassabis would one day embark on.

It’s a significant recurring anecdote because he uses it to accentuate his humility — or even his humanity. He was just a little boy. And yet he traveled the world going up against adults. An exceptional little boy, then. It is also a reminder that his ultimate inspiration is the human brain, its sanctity paramount to both Hassabis’s research and his self-image. Elsewhere, after all, he posits that the development of superhuman AI will reveal the secrets of nature and so bring man closer to God. “For me,” he tells Mallaby, “science is a spiritual endeavor”.

There are striking echoes of history here. Until the dawn of modernity and secularism, political figures had to insist on their ultimate servitude to God even as they sought hard power in the temporal world. Thomas Cromwell, in his famous portrait by Holbein, appears in the sober garb of a serious power-player — one who has just orchestrated the submission of England’s clergy to King Henry VIII, fundamentally changing the shape of religion in England forever. Yet a Book of Hours sits on the table in front of him, signaling his continued devotion.

If we are to believe predictions, AI will have a far greater effect on our society than did the Dissolution of the Monasteries. Such a change cannot be forced through without backlash. To mediate it, today’s AI bosses must project respect for the human, even as they build something that could surpass us. As Cromwell doffed his cap to faith, even while ushering in shattering changes to religion, his Silicon Valley successors must position their work as serving humanity while upending daily life. Hassabis has repeatedly demonstrated this respect, idolizing the human brain through his neuroscience research and talking of a quasi-religious calling to unlock the secrets of life.

To figure out what’s really driving his protagonist, Mallaby is particularly concerned with what Hassabis and his ilk are reading. His ignorance of the books of the AI prophet Ray Kurzweil sets Hassabis apart from other attendees at the Peter Thiel-backed Singularity Summit, a hive of the early AI scene. Of far greater literary interest to Hassabis is the eponymous protagonist in Ender’s Game. The story of a gifted child tasked with saving humanity resonated so deeply with Hassabis that he asked Mallaby to read it in advance of their first interview. Elsewhere, great scientists are his chosen models. “Isaac Newton, Alan Turing, all my heroes,” Hassabis says of walking down a street in Cambridge. “I could feel them in the bones of the stone, their intellect and vision. They were almost calling out to me.” He is often getting goosebumps like this, taking pictures of the Nobel Foundation’s guestbook signed by Einstein and Feynman. It is charming, suggestive of the eager, nerdy young boy under the surface. But it is also another way of self-fashioning, absorbing those figures into his own story. Here they stood and now I am here too.

Hassabis also defines himself against others, especially OpenAI’s Sam Altman: “I’m doing it for knowledge and science. He seems like he’s doing it for power.” He seems a little baffled by Altman, who comes from a new mold. Instead of Hassabis’s deference to the God-like power of nature, Altman starts to build a cult-like following for himself.

This all makes me wonder how much of the access granted for the book, and the production of the DeepMind documentary, The Thinking Game, and the appearances at world conferences and in interviews, is actually about Hassabis craving recognition. Altman, remember, not only kicked off an AI race by releasing ChatGPT, but in doing so made himself the defining figure of AI.

Suggested reading

Inside the AI resistance

By Farahn Morgan

Admittedly, some of Hassabis’s efforts to build his public profile predate the release of ChatGPT. But the invisible hero narrative feeds back into those twin competing pillars of exceptionalism and humility on which Hassabis perches his identity. Mallaby allows his subject to speak at length in the final pages, where Hassabis sets out his own vision of his place in the world. “I’m a weird British outlier, on this little island here, and I’ve made my own path. I’ve followed my passions and tried to stay true to what I believe in.” His refusal to leave London, for which he gets a lot of credit in British policy circles, becomes part of a narrative of being not like the other tech bros.

It should hardly be surprising that someone who spends much of his time thinking about consciousness, both human and artificial, might be particularly preoccupied with his own identity — and how that should be projected. “In my world, humanism and spiritualism and science all go together,” he says. And, like Cromwell, it is to a Renaissance figure that he looks for a model. “Da Vinci is my favorite because everything’s just flowing into one river. And that’s how I try to live.” The meaning is obscure, but Hassabis seems less enthusiastic to incorporate that element of his personality which thrives on competition and the exercising of power —though he is willing to accept them when Mallaby points them out. “I have to do what’s necessary.”

Hassabis, in short, appears to be banking on a longer-term legacy, one that emerges long after the machinations of the court are concluded. But if he or any of his rivals actually succeeds in creating AGI, this careful self-fashioning may become less relevant. If you change the world, it won’t be the books or paintings that dictate your reputation, but the consequences.

Since the mid-20th century Americans have distinguished the “upper middle class” from the “lower middle class.” We can think of these as two lifestyle packages, combining certain jobs, certain homes and neighborhoods, and certain family structures.

The upper-middle-class lifestyle package has three elements: a well-paying job that requires a college diploma, preferably in a large, high-status institution, along with a trophy house in a prestigious neighborhood that functions both as a display room for fancy furniture and a speculative asset. Another part of the upper-middle-class package has been the ability of affluent professionals to delegate care for children and elderly relatives to servants or paid employees of institutions (public, private, or nonprofit).  

More from this author

The myth of Western crisis

By Michael Lind

Contrast this with the lower-middle-class lifestyle package: a decently-paid job that requires only a high school diploma or some additional training and sometimes a bachelor’s degree; a house that is productive; and reliance for care-giving chiefly on family members, supplemented sometimes by neighbors or friends. The productive household might generate some income by being the site of a small business or, with the help of appliances, allow residents to substitute unpaid DIY labor for market transactions, by building, fixing and refitting at home. 

The mass middle class that was created in the United States after World War II was a mass lower middle class, not a mass upper middle class. There were various paths of entry into the mass lower middle class, whose members included unionized workers, prosperous skilled artisans, minor functionaries, small business owners and even well-off family farmers.

Neither a college education nor a professional career was required for entrance into the mass lower middle class, between 1945 and the 1980s. In 1960, only 7.7% of Americans had bachelor’s or higher degrees; by 1980, around the time I graduated with a BA, the number was still only 16.2%.  

Between World War II and the last quarter of the 20th century, most Americans graduated from high school and went to work and got married and had children. Being part of the middle class was defined not by educational or job status, but by a certain material standard of living, like having a family home, at least one car, and a television that received the three commercial network channels (created in 1970, PBS was limited to big cities and college towns).  

Beginning in the Reagan and Clinton years, however, the American Dream was redefined to mean achieving the lifestyle of the small upper middle class — not the super rich, but the most affluent lawyers, doctors, bank executives, corporate executives, and tenured professors. We were supposed to believe that their affluence was the result of their superior skill and merit-accruing academic labor, as measured by tests and rewarded with graduate and professional degrees. And it was true that brain power counted for something. But just as important was the success of the legal, medical, and academic professions a century ago in the 1900s in creating professional associations like the American Bar Association (ABA) and American Medical Association (AMA), which captured occupational licensing in their fields and created cartels that protect their members from fee-lowering competition. 

We were also supposed to believe that upward mobility was achievable en masse, through a college diploma. By the 1990s, politicians, corporate executives, nonprofit reformers, and self-interested academic bureaucrats united behind this message. And Americans listened. The number with a BA or higher has exploded from 16.2%  in 1980 to 38.7% in 2024. In some European countries the college-graduate share of the population is even higher.

Suggested reading

The plot to replace teachers with tech

By John Allen Wooden

But the higher-ed industry was “talking its book,” i.e. painting a biased and overly rosy picture, when it sought to expand the number of applicants. And business leaders were pretending that stagnant wages were the result of inadequate education and insufficient skills on the part of workers, rather than the offshoring of industry, the flooding of the labor market with cheap immigrant labor, and the near-annihilation of private-sector labor unions. For their part, politicians of all persuasions found that encouraging more Americans to go to college cost little and avoided controversy.

The result of this massive and ill-advised expansion of higher education in the last half century has not been upward mobility, but too many college graduates chasing too few well-paid professional jobs. In 2025, only 19.3% of job postings in the US required a bachelor’s degree or higher, at a time when those qualifications are possessed by around 40% of the population. According to the St. Louis Fed, a year after graduation, 52% of Americans with college degrees are employed in occupations which do not require their degrees; the number after 10 years is 45%. For around half or more of college graduates, then, investing in a college diploma has been a bad decision.  

What about the second part of the upper-middle-class lifestyle package — the big trophy house in a prestigious neighborhood? There are plenty of modest houses in unfashionable neighborhoods, cities, and regions. But the Starbucks barista with a useless degree in communications or marketing wants to live in a hip neighborhood in New York or San Francisco or Austin, not in a tract-home neighborhood in Cleveland, Ohio.

The final element of the upper-middle-class lifestyle package is outsourcing the caretaking and domestic roles to unrelated paid caregivers. The rich have always been able to afford to hand off their children and elderly to be looked after by servants or other paid non-relatives, and to delegate household chores to other servants and service workers. Today, however, there’s a “care crisis” in the form of college-educated professionals who belong to two-earner couples, who can’t afford private child care and private schooling. The rich can afford to be generous to Mary Poppins or Jeeves. But the struggling two-earner professional couple typically relies on poorly-paid, largely immigrant female labor to look after the children and keep the house or condo clean.

The point is this: half or more of the college-educated professionals in the US belong to a pseudo-upper middle class, a knock-off version of the actual professional-managerial elite.  These wannabes and also-rans have worthless diplomas and jobs which, however honorable in themselves, do not allow them to purchase the homes and lifestyles of which they dream. They have champagne tastes on beer budgets.

This ersatz upper middle class has been overbuilt from the late 20th century until the present. It was unsustainable and bound to collapse. By automating a lot of “knowledge jobs,” AI may accelerate the inevitable contraction. But the decline was bound to come at some point even without AI.

If the half-century experiment in creating a mass upper middle class of college-educated professionals has failed in the US and similar democracies, what is the condition of the mass lower middle class? Current data prove that lower-middle-class neighborhoods and lifestyles, far from being gone with the wind, continue to be as distinct from the professional class as they are numerically significant. 

Working-class people of all races are still more likely to be family caregivers or to rely on family caregivers than to rely on paid nannies or institutions. They are likely to live in multigenerational families, a trend that is growing. They have side-hustles and run home-business — in fact half of small businesses are run from a home or began there. And working-class and low-income people are more likely to engage in home improvement projects than the affluent, who hire specialized workers instead. According to a survey last October, more than half of American parents said they would make Halloween costumes for their children from scratch, while four in ten — presumably more upscale — said they would buy them.

For members of today’s bloated and unsustainable college-educated professional class who identify their own real or desired lifestyles with success, all of these traits of what I am calling the lower middle class represent the failure of American society. How terrible, that some children receive low-quality day-care from grandparents or aunts or uncles, instead of from an immigrant nanny or corporate daycare workers or public school teachers. How tragic, to think that some middle-aged people must let elderly parents live with them because they can’t afford to warehouse them until they need round-the-clock nursing care in an expensive, high-quality retirement community. It’s a national disgrace that some high-school graduates still run some small businesses out of their homes, when they might have been salaried college-educated professionals with their own cubicles in office buildings somewhere. Imagine, there are still Americans so deprived that they make the costumes that their children wear at Halloween, instead of buying them in stores. 

But the lower middle class may have the last laugh. AI may well wipe out whole chunks of the upper middle class. But innovative technologies may not only spare many people in the lower middle class, they might also empower some.  

Computers soon may be able to write wills and diagnose patients and grade papers, causing a mass extinction event for lawyers and doctors and professors. But it will be some time before physical robots are able to replace the “essential workers” who were lauded during the COVID pandemic — nurses, caregivers, first responders, police officers, public and private utility engineers and mechanics.

If you are an executive vice-president of a firm whose work consists largely of convening meetings and sending emails, you may soon find yourself the victim of technological unemployment. But if you run a cookie-baking business from your home, you can already do much of your accounting and tax work with software, while reaching a large client base online. Advances in AI will only accelerate these possibilities.  

If you are a middle manager in a large corporation, your job may vanish when many of those you supervise are replaced by computer programs. But if you like to tinker in your garage, instead of doing woodwork with old-fashioned tools you may be able to use 3-D printing and other advanced technologies to make toys, models, artwork, dishes and cutlery, household furniture, musical instruments, and even functional drones.

If you are a banker who evaluates loan applications, you may be replaced by bots that do the same job in less time and without human error. But if you want to rent out a spare bedroom, or rent an unneeded space in your two-car garage, you can do so with the help of Airbnb or other online services.

If you are an academic who specializes in researching archives, your occupation may not be long for this world. But if you want to home-school your children, technology will permit them to take virtual classes, listen to virtual lectures, and have access to first-rate libraries, while learning social skills by playing with neighborhood kids.

Suggested reading

We are all Eurasians now

By Antaol Lieven

If your job as an insurance company apparatchik is to boost company profits by denying coverage to your company’s paying clients, software can cheat sick people just as well. Meanwhile, telemedicine for doctors’ appointments, home visits by nurses, and home-based monitors connected to hospital or medical practice computers might allow your ailing, elderly parent to live with you in your home, instead of being committed to impersonal institutional care.

The obstacles to many of these classic lower-middle-class activities are not technological but legal. Many technology-enabled opportunities can be realized only by loosening or repealing restrictive local government and zoning regulations that bar businesses from being run out of homes, ban accessory dwelling units (ADUs) or “granny flats,” and prevent homeowners from renting rooms to non-relatives.  

Today’s YIMBY (Yes in My Back Yard) movement mostly represents the interests of greedy real-estate developers who want to plop down apartments in desirable single-family neighborhoods, and of social-climbing, childless singles willing to be packed like sardines in hip neighborhoods if they can enjoy a cool urban consumer experience in their 20s. What is needed is not replacing single-family neighborhoods with blocks of apartment buildings, but rather eliminating restrictions which prevent sterile, snooty upper-middle-class neighborhoods from being converted into dynamic lower-middle-class neighborhoods, with home businesses and garage craftshops and maybe the occasional chicken in the backyard.

Needless to say, property owners in upper-middle-class neighborhoods would fight these reforms.  But there are not enough affluent young professionals in the next generation who can afford the upper-middle-class lifestyle, and there would not be even if the progress of AI turns out to be exaggerated. It is conceivable, of course, that sufficiently advanced technology might lead to truly widespread unemployment or lower wages in most occupations and income levels. But the idea that there is a binary choice between a well-paid job and a universal dole reflects the bias of the professional class.  

In a lower middle class or working household, there may be multiple streams of income. There are full-time jobs for some members of the family, side hustles, and rental income, to say nothing of DIY projects that save money, compared to purchases of finished goods or labor. Meanwhile, government may also subsidize a multigenerational household in the form of child allowances and Social Security, as well as means-tested in-kind benefits like food stamps if necessary. In the event of a true robotic jobs apocalypse, it would be easier to incrementally increase these existing government cash and in-kind subsidies to households, in addition to their other income streams, than to provide everyone with a cash stipend or “universal capitalist” investments in stocks and bonds like Tony Blair’s failed “baby bonds” and the doomed-to-fail “Trump Accounts.” 

More from this author

MAGA’s samurai justice in Venezuela

By Michael Lind

In hindsight, the US and similar industrial democracies took a wrong turn after the 1970s when “middle class” was redefined to mean “upper middle class” rather than “lower middle class.” There is no reason why the post-1945 lower middle class, rather than the college-educated professional class, could not have continued to define the middle class to this day. The silicon chip did not force Western democracies to shower expensive college diplomas on great numbers of citizens who got jobs that do not require them. The internet did not compel people to aspire to move from lively, eclectic lower-middle-class neighborhoods to sterile upscale bedroom communities policed by authoritarian homeowners’ associations. Satellite telephony did not persuade people to equate social status with sitting in an office cubicle emailing memos instead of running a business from the garage or doing DIY projects at home. The idea that from the age of 18 until death in old age, the members of a multigenerational family should all have separate homes was not suggested by a computer program. Far from being determined by technology or economic trends, the failed experiment in ballooning the upper middle class in the last half century was the result of changing fashions, misguided reforms, and the self-aggrandizement of professionals as a group.

Having spent portions of my life in both the lower middle class and the upper middle class, I find the lower-middle-class environment more … well, more human. Born in 1962, I spent the first year of my life in a classic lower-middle-class, blue-collar neighborhood on the east side of Austin, Texas, which has now been gentrified beyond recognition. As an infant I lived with my parents in a studio apartment atop the detached garage behind the small house of my widowed paternal grandmother, a bank teller who made additional income by renting one of the two bedrooms in her home to female University of Texas students. My college-educated parents later moved the family to suburban houses, first in South Austin and then in West Austin. But our family spent a lot of time with my grandmother in East Austin and I lived with her as an undergraduate at UT-Austin.

My grandmother’s lower-middle-class neighborhood was like a tropical savanna teeming with wildlife, compared to the Arctic sterility of a professional-class suburb. One of my grandmother’s neighbors, a high school teacher, spent hours with the woodworking equipment in his garage. Another neighbor was a black lay preacher who supported his prison ministry by mowing lawns. A few doors down was a ladies’ “beauty parlor,” possibly unlicensed, that was operated by yet another neighbor out of her small home. Around the corner was a small grocery store owned by a relative, and one block away was a service station owned by a family friend, where I spent pleasant hours as a child watching garage mechanics repair cars, as my paternal grandfather, who died before I was born, had done. Two blocks away was the black-owned and staffed barbershop where my father in the 1960s took me on occasion to get my Bart Simpson-like buzz cut.

If there was a homeowners association with restrictive rules in my grandmother’s lower-middle-class neighborhood, it was moribund in the 1960s and 1970s. People ran nursery schools out of their homes and rented rooms and put up cars on concrete blocks in the front yard and kept hens and crowing roosters decades before hipsters embraced backyard poultry. In an elite Austin suburb, if you ran a business out of your home or rented a spare bedroom, your neighbors, obsessed with possible declines in the value of their trophy houses as a result of the “wrong sort” coming in and out of the neighborhood, might sue you.

I for one will not lament the shrinking of the oversized upper middle class in relation to the lower middle class. When vacant downtown office towers are detonated and depopulated, university campuses are rented or sold off, and pretentious McMansions rot in the rain for want of buyers, humble lower-middle-class neighborhoods and their denizens will survive — and perhaps even flourish, with the help of advanced technology.  

In a post on X, Reichstadter explained: “I am calling on Anthropic’s management, directors and employees to immediately stop their reckless actions which are harming our society and to work to remediate the harm that has already been done.” Days later, filmmaker and former AI researcher Michaël Trazzi launched his own campaign, inspired by Reichstadter’s, outside Google DeepMind’s London headquarters.

The cross-continental demonstration took hold, earning press coverage in Business Insider, Futurism, and The Verge, among others, even if Anthropic and DeepMind themselves did not respond. After 30 days of consuming only vitamins and electrolytes, Reichstadter ended his participation in the protest. “We who have everything at stake,” he wrote, “cannot stand by while this malignant conspiracy against the safety and security of the human race flourishes under the aid and protection of the corrupt and inept governments of this world.”

***

In the months after the hunger strike, I connected with Reichstadter through some mutual acquaintances on X. I was told that he was part of a larger anti-AI social movement cropping up in San Francisco, and that he had helped found an organization called Stop AI. The city is at the center of the AI boom — and corresponding safety movement — but it’s also the beating heart of American counterculture. I wondered what this burgeoning social resistance might look like, so ask Reichstadter to meet.

He agrees, and we make plans to go hiking in the Berkeley Hills. Given Reichstadter’s past, the natural setting feels appropriate. With a background in climate and other issue-driven activism, in April 2022 he went on a two-week hunger strike outside the offices of Miami-Dade County Mayor Daniella Levine Cava to bring attention to climate change. In June that same year, the father of two spent 28 hours atop the Frederick Douglass Memorial Bridge in Washington, D.C. to protest the Supreme Court decision overturning Roe v. Wade. “Civil resistance works,” he said in an Instagram video facing out over the Anacostia River. “Learn about it… That’s how we’re going to get the radical change we need in the shortest time necessary.”

More from this author

The small-town Alabamians fighting a data centre

By Farahn Morgan

Out on the trail, Reichstadter walks me through his journey into anti-AI activism. He says he was still in college when it first occurred to him that super-powerful AI could cause serious human harm. With the launch of OpenAI’s large language model ChatGPT in 2022, his fears became more concrete. Now, as the technology continues to advance at breakneck speed, Reichstadter’s efforts have become increasingly urgent. “These systems,” he tells me, “have the potential to become powerful enough, even without consciousness, to control the world.”

Reichstadter first came to San Francisco in 2024, after meeting a guy named Sam Kirchner through a group called Pause AI. The pair apparently bonded over the idea of using civil resistance and direct action, which Pause AI does not allow, to bring attention to the risks surrounding the development of super-intelligent AI. That led to a new group: Stop AI. Reichstadter is no longer connected with it — he left last August amid internal disputes — but still maintains ties: including legal ones. He’s awaiting trial with another Stop AI member for charges related to non-violent protests last February at Open AI’s headquarters. Lately, Reichstadter has been acting independently and recruiting more people for the cause. “This is not necessarily about reaching everybody,” he says. “It’s about finding a viable coalition of individuals who are willing to push back.”

***

The idea that superintelligent AI might lead to catastrophic harm or human extinction, once the domain of theorists and philosophers, has gained momentum among some experts. The discourse surrounding the extinction threat — or “x-risk” as it’s known in some corners of the internet — became especially intense following the release of ChatGPT. More than 1,000 tech leaders and experts, including Elon Musk, signed a March 2023 letter calling for tech companies to pause further AI development. In May the same year, Nobel laureate and pioneering “AI godfather” Geoffrey Hinton dramatically left his job at Google to warn about the dangers associated with the rapid development of artificial intelligence. Weeks later, CEOs from leading AI labs, including Anthropic, DeepMind, and OpenAI, acknowledged that the extinction risk from AI is on par with that of pandemics or nuclear war.

Experts disagree about what the actual risk might look like, and how likely any given scenario might be. One argument suggests that superintelligent AI systems pose a risk to human beings if their goals are not precisely aligned with human values. The idea is demonstrated by Nick Bostrom’s “paperclip maximizer” thought experiment, in which a powerful AGI is given the directive to maximize the number of paperclips in existence: even if that means eliminating human beings in the process. Others worry that bad actors could misuse AI to cause more conventional harms, from global war and disease to economic and ecological collapse.

Even so, big tech firms have continued full speed ahead toward increasingly powerful AI systems, with critics ever-more worried in turn. Stuart Russell, a professor of Computer Science at UC Berkeley and one of the world’s leading researchers in artificial intelligence, has likened the situation to an “arms race”. “For governments to allow private entities to essentially play Russian roulette with every human being on earth is, in my view, a total dereliction of duty,” he told AFP during last week’s India AI Impact Summit. The warning comes as researchers at some of the world’s biggest AI firms have also tendered their resignations. As Mrinank Sharma, who led Anthropic’s safeguards research team, wrote in an open letter to his colleagues: “The world is in peril. And not just from AI, or bioweapons, but from a whole series of interconnected crises unfolding in this very moment.”

***

In some ways, Stop AI is an effort to harness those anxieties into collective action. It’s an approach that’s been used by radical environmentalist groups in the UK, notably Extinction Rebellion and Just Stop Oil, to bring about systemic change. Traditionally, those within the AI safety movement have taken the insider approach, working with labs and research centers and think tanks to influence safety procedures and government policy. That, along with concerns about being taken seriously, have encouraged caution when it comes to practical campaigning. Stop AI, however, explicitly favors civil disobedience. “History has shown us that non-violent movements can meet their goals quickly when a large enough group of people cooperate towards a common goal,” reads the group’s website. “Non-violent resistance is the only answer.”

As its name would suggest, Stop AI’s end goal is a permanent and binding international ban on the development of both Artificial General Intelligence (AGI) and Artificial Super Intelligence (ASI): respectively AI as capable as human beings at most cognitive tasks, and AI that surpasses the abilities of human beings. In its pursuit, the group has never shied away from controversy, adopting tactics like blocking entrances to office buildings or chaining their doors shut.

Last November, they made national headlines when a public investigator hopped on stage to serve a subpoena on OpenAI CEO Sam Altman during an event with Golden State Warriors coach Steve Kerr. On social media, Stop AI claimed responsibility for the stunt. “Our public defender successfully subpoenaed Sam Altman to appear at our trial where we will be tried for non-violently blocking the front door of OpenAI on multiple occasions,” the group said in an X post. “This trial will be the first time in human history where a jury of normal people are asked about the extinction threat that AI poses to humanity.”

In the past, the group’s messaging has been direct, if not apocalyptic. There was a banner, for instance, that read “AI Will Kill Us All.” Then there was the accompanying chant: “Stop AI or We’re All Gonna Die.” In recent months, though, as Stop AI’s organizers have wrestled with how to more effectively raise awareness and grow their ranks, that message has evolved. With the election last fall of new leader Matthew Hall, who goes by the name Yakko, Stop AI began moving from focusing solely on the risk of human extinction. “I feel a moral obligation to warn people,” Yakko told The Daily Californian. “I just think when people get too scared they just kind of shut down. They still have to go about their daily business.”

What happened next hints at the broader tensions within the anti-AI movement. Kirchner, apparently frustrated by what he saw as a departure from Stop AI’s more radical message, stormed out of a November planning meeting. Later, he came back and asked for access to the organization’s funds. Agitated that his request was denied, he struck Yakko several times in the head. According to Yakko, Kirchner told him that “the non-violence ship has sailed for me”. He later apologized, but the incident led to his expulsion from Stop AI. Since then, though, Kirchner has vanished, not turning up for a scheduled court date. His bike and camping gear have disappeared from his house too. Out of an abundance of caution, a member of Stop AI called the police to say that Kirchner could potentially pose a threat to people working at AI labs. That led to lockdowns at OpenAI, though members of Stop AI maintain they are not aware of any specific threats.

Either way, StopAI was quick to post a message on X detailing the situation, reaffirming its commitment to non-violence. Pause AI equally condemned Kirchner’s actions, explicitly stressing that it “has no affiliation with him or his organization”. For its part, Pause AI’s Code of Conduct explicitly prohibits violence, while also requiring compliance with any applicable laws. Pause AI CEO Maxime Fournes adds that the movement has always been “extremely peaceful”. That’s echoed by a broader debate about the perils of a certain kind of apocalyptic anti-AI messaging. “An early activist at Stop AI had a mental health crisis and went missing,” Remmelt Ellen, an advisor to Stop AI, warned in an X post. “Act with care. Find Sam. Stop with the ‘AGI may kill us all by 2027’ shit please.”

It’s impossible now to know what Kirchner might have been thinking, what pressures he may have been under. Yet three months after his initial disappearance, the 27-year-old activist is still missing. According to Kirchner’s friends and allies, he had been under a lot of stress, and the thought that AI might cause harm to his sister was consuming him. “He had the weight of the world on his shoulders,” one of the group’s members told reporters.

Lately, Stop AI has been engaged in the delicate task of sorting out how to continue its work in spite of everything that’s happened — and despite the fact that Kirchner has still not been found. For its last public event of 2025, Stop AI hosted what it called “A Vigil for Humanity”. Tellingly, perhaps, it wasn’t held at OpenAI or Anthropic, but rather at San Francisco City Hall.

The tone was decidedly somber. A video posted online shows members holding lanterns in the waning daylight. Someone in the group explains who Stop AI is and what they’re doing there. There are warnings about the extinction risk. But then the conversation turns toward the broader value of human life. “We are going to take a moment to meditate and pray that every human life is protected, including the lives of the people who work at these companies… even though we deeply disagree with what they’re choosing to dedicate their lives to… even though they’re pursuing something that is dangerous. We think their lives are sacred and important too.”

***

While in San Francisco, I planned to go to a Stop AI protest at OpenAI’s headquarters, but then learned that the event had been canceled. I considered going by the offices anyway, just to get a sense of the place. Instead, I connect with Wynd Kaufmyn, who is one of the group’s organizers. We meet at a restaurant near the Fruitvale BART station; she’s planning to attend a nearby protest against the Trump’s administration’s immigration policy.

Now in her sixties, Kaufmyn came to the Bay Area as a Berkeley graduate student — drawn in by its radical politics and the lingering energy of the Free Speech Movement — and never left. She met Reichstadter and Kirchner at a protest in support of Palestine at Andrews Airforce Base. The pair shared some information about Stop AI, and Kaufmyn started going to meetings. Now, more than a year later, she’s a member of the group’s leadership council. “The reason that we’re having this AI problem right now is that we have corporations running the world,” she tells me. “And they don’t care what happens to people. So, for me, this is really about being anti-establishment, being anti-fascist.”

Stop AI, Kaufmyn explains, has been talking a lot more lately about AI skepticism as a big-tent issue, encompassing concerns about more immediate risks — like autonomous weaponry, surveillance, mass resource use, and job loss — alongside existential risk. “We want to be honest about what’s at stake here,” she says. “But we don’t want the message to be all doom.”

Certainly, Kaufmyn worries that anxiety about impending catastrophe is what led to Kirchner’s disappearance. “I could tell he was really depressed,” she says. “And I wanted to help him, but he was always very guarded.” She echoes the observation that he had become increasingly frantic that AI might hurt or kill his sister. “There was an article in the San Francisco Standard that I thought put it well,” Kaufmyn continues. “Is it possible to protest the end of the world without losing your mind?”

Suggested Reading

Can we escape AI slavery?

By Yanis Varoufakis

In recent months, Stop AI has been in a regrouping phase — trying to understand what happened and how to keep it from ever happening again. Kaufmyn explains they’ve been working on strengthening the organization from the inside out, building a strong internal culture so that they can reach more people outside of the movement. They’re being mentored, she tells me, by some members of Extinction Rebellion. The group is known for practicing what it calls “regenerative culture”, which tries to create healthy relationships through mutual care. But Extinction Rebellion is also among those in the environmental space most willing to aggressively engage on the AI issue.

Last summer, two Extinction Rebellion activists protested Meta’s AI expansion by chalk-spraying the windows at its New York offices with messages like “Meta Makes Kings” and “Zuck Loves Trump”. Earlier this month, Reichstadter shared an X post from Extinction Rebellion Founder Roger Hallam, announcing the Zoom launch of a group called “Pull the Plug” in the UK. “Give ordinary people a say in how AI is used in our lives,” its website says. Later this month, the group is hosting what’s been billed as the UK’s “first-ever anti-AI march” at OpenAI’s London offices. It’s called the “March Against the Machines”.

Before I leave San Francisco, I tag along with Kaufmyn for part of the immigration policy protest. It’s crowded in Fruitvale Commons. Some of the activists shout over a megaphone. Others hold a huge white sign with red lettering that says: “People Over Billionaires.” Kaufmyn surveys the banner and reads the words out loud. “People over billionaires… I like that.” For a moment, we stand in silence. “I have one last question,” I tell her. “What do you think your goal is for Stop AI?” She shifts her weight a little. “I would like everybody from every walk of life to stand up to these billionaires — through civil disobedience, a hunger strike, however — and demand that they aren’t building harmful technologies that could destroy our lives when we are telling them not to.” A different battle in the same fight within the AI-skeptic movement, and one Kaufmyn and others have fought many times before.

The rise of AI, the so-called Fourth Industrial Revolution, is now reviving that old dread, but this time it portends a far more volatile political horizon than the last one. It’s now the college-educated, office-dwelling winners of the age of the internet and globalization, who are suddenly discovering that they, too, can become the needless things alongside the blue-collar manufacturing base that Trump and MAGA has thus far failed to revive. 

We will soon find out which way the political winds will shift in response. The first act of the AI revolution was a content explosion. Generative AI and large language models like ChatGPT, Claude, and Google Gemini flooded the internet in the early 2020s with cheap words, cheap images, and cheap expertise, warping media, education, and the economy of attention into something closer to a hall of mirrors. That phase began bending politics by accelerating misinformation, flattening expertise, and lowering the costs of persuasion. We are now entering a more destabilizing phase: job loss.

As of 2026, AI “coming for jobs” is no longer a speculative future-of-work panel at Davos. In a limited sense, it’s already here. By the end of 2025, AI had been cited as a contributing factor in nearly 55,000 domestic layoffs, according to executive-coaching firm Challenger, Gray & Christmas. The January jobs report was rosier than expected, but the underlying data was grim. Healthcare and social assistance accounted for 124,000 of the 130,000 jobs gained last month, industries essentially dedicated to managing an aging, poorer population. An AI executive at a top consulting firm told me last week that end-of-year updates — Opus 4.5, the Anthropic model that launched at the end of November; and GPT-5.2, the OpenAI model that launched in December — have crossed a threshold and can be trusted to handle more complex tasks with fewer mistakes. “Most people don’t realize it yet. Now [AI] is at the level where I think we’ll start seeing real economic impacts this year, and it will snowball relatively fast,” she told me. Investors have noticed, and software stocks are being punished on Wall Street as markets begin to price in the possibility that AI handles the bulk of global programming.

The snowball may have already begun rolling. In January, American employers announced more than 108,000 job cuts, the worst start to a year since the 2009 Great Recession, with layoffs up 118% year over year and up more than 200% since late 2025. The casualties are not obscure startups. Amazon eliminated roughly 16,000 corporate jobs in a single month as part of a broader plan to trim tens of thousands of white-collar roles. Meta pruned 1,000 jobs from its virtual reality unit Reality Labs. Salesforce’s CEO acknowledged that the company shed 4,000 customer-support workers after AI tools absorbed roughly half of the workload. What we’re seeing now isn’t mass unemployment but a subtler erosion: fewer entry points, thinner career ladders, and a growing sense that the white-collar future once sold as stable and dignified is about to go the way of the video store clerk. 

More from this author

The mutual-aid delusion

By Ryan Zickgraf

It’s about to get worse: CEOs aren’t hiding it anymore; they’re now openly advocating reorganizing firms around AI labor as a cost-cutting imperative. Dario Amodei, the CEO of the AI company Anthropic, said that AI could drive unemployment up 10 to 20% in the next one to five years and “wipe out half of all entry-level white-collar jobs.” Microsoft AI CEO Mustafa Suleyman is sounding an even louder alarm, predicting that most white-collar work “will be fully automated by an AI within the next 12 to 18 months.” Progressives like MSNBC’s Chris Hayes are calling this development a deliberate class war by Big Tech CEOs against white-collar workers for being too uppity, but this is not a conspiracy against professionals. It is capitalism working exactly as intended. The software doesn’t have to be perfect; if it can get to “good enough” at drafting reports, writing code, moderating content, summarizing documents, or optimizing logistics, then the $120,000-a-year credentialed professional becomes extinct, as more profit can be extracted. Robots and conveyor belts did this to factories and will continue to do so; now models and chatbots are doing it to the cognitive economy, turning thinking into the next assembly line.

This will upend politics because jobs lie at the heart of America’s social contract. As Franklin Roosevelt warned a century ago, “people who are hungry and out of a job are the stuff of which dictatorships are made.” We have already seen how downward mobility reshapes politics because we’re still feeling the tremors of the last employment-quake. In the 50 years since Terkel’s Working, deindustrialized and job-starved blue-collar communities lost more than jobs. Factory closures unraveled unions, apprenticeships, and entire local economies, shunting displaced workers into a service sector defined by precarity and downward mobility. Elites sold the transition as inevitable progress and told blue-collar workers to “learn to code,” as if your uncle could simply download a new career. The promised “knowledge economy” absorbed a narrow slice of the workforce, leaving the rest to navigate wage stagnation, disability rolls, and social decay in towns that had lost their reason to exist. Over time, this produced institutional estrangement: trust in government, the media, and expertise waned; participation declined; and resentment grew. 

And so, the anxiety born of the previous wave of automation and the legitimate economic grievances of blue-collar communities were largely rerouted into a new political identity, one that fused material loss with cultural resentment. As unions withered and Democrats recast themselves as the party of the ascendant professional classes, many working-class voters concluded that no one in power was actually trying to rebuild the world they had lost. Into that vacuum stepped MAGA and a Right-wing populism that promised results but has instead given rise to a politics of recognition — an inverse woke — embodied in the unlikely avatar of Donald Trump.

So, what happens to American politics when the script is flipped, and we enter a new era of white-collar precarity? We can look back to the recent past and recall that, after the 2008 recession, it was young men who got especially angry. Downwardly mobile urban millennials drifted toward radical Left-wing politics, including the Occupy Wall Street movement and both Sanders campaigns, myself included. In the current decade, the Gen-Z men shut out by elite institutions often join their grandfathers and turn toward MAGA, or worse, into Groypers. But an AI-driven white-collar apocalypse has no equivalent of the American Rescue Plan around the corner, and it will move faster through institutions because the people experiencing it — journalists, lawyers, policy staffers — are the ones who produce political legitimacy itself. When that class loses faith in the system’s stability, the political climate may quickly become volatile. 

Marx, the prophet of the Second Industrial Revolution, would argue that as economic obsolescence metastasizes into political instability, a revolution will soon be upon us. After all, yesterday’s bourgeoisie is soon to become tomorrow’s proletariat. That’s unlikely, but if AI brings about a college-educated jobs apocalypse rather than a utopia of zero work and universal high income, then we will see a more drastic political shock over the next decade than we’ve seen in our lifetimes. Call it W.H. Auden for the white-collar age. “The Age of Anxiety” was the poet’s name for a post-WWII world in which the old guarantees had collapsed, and everyone felt one missed-step away from the void: “the wolves will get you if the moths won’t,” he warned. Large catastrophe or nibbling decay: either way, your previous life is over.  

In the short term — the next one to three years — the primary victim of this upheaval will inevitably be the Republican party and the MAGA movement. Economy and jobs remain the top issues for American voters. Whether it is fully deserved or not, the electorate will blame the incumbent leadership for the AI-driven displacement. The current administration, having tethered its brand to a return to American greatness and a blue-collar resurgence, will find itself presiding over a white-collar bloodbath. The 2026 midterms are shaping up to be a massacre for the GOP as voters react to tariff-driven inflation and rising costs for everyone. By 2030, both parties will face the wrath of young people who are expected to build a life on internships, side hustles, and selling themselves on OnlyFans. 

But this isn’t necessarily a victory for the center-Left or center-Right. Mainstream democrats are trapped between worker-friendly rhetoric and donor-class dependence on the industries driving automation. Republicans are trapped between populist grievance, pro-market orthodoxy, and vying for Big Tech’s hand in political marriage. The lesson that the pandemic-era satire Eddington expertly captured was this pathology: social media trained Americans to see their neighbors as existential enemies based on voting behavior, while the political establishment quietly benefited from the infrastructure boom, tax abatements, and data-center gold rush that accompanied the platform economy. The culture war supplied the noise, and capital captured the signal.

But maybe not forever. We are already seeing the emergence of a new, non-partisan “AI Populism” where something like the “Humanists” square off against “Tech Accelerationists.” Across the country, local communities are already revolting against the construction of massive AI data centers, even without the sting of massive job loss. These eerie windowless temples of compute strain local power grids and suck up millions of gallons of water, and arrive with the footprint of a fleet of spaceships: alien, resource-hungry, and indifferent to local life. The political backlash is no joke. These fights are erupting in swing states and battleground districts, animating local elections and straining party coalitions. 

Suggested reading

How Big Tech killed literary culture

By Nicholas Carr

In Pennsylvania this week, the Montour County commission denied rezoning for a proposed data center tied to a major tech company after residents warned it would hike utility bills and worsen energy insecurity in a region already wary of industrial disruption. In Arizona, grassroots groups gathered enough signatures to put data center siting decisions on the ballot, signaling that tech infrastructure could soon be one of the few issues that actually mobilize voters outside the traditional culture wars. These skirmishes are shaping up to be an early indicator of what the next phase of the AI revolution will look like politically. Not a clean Left-Right realignment, but a messy conflict between those who bear the costs of automation and those who harvest its efficiencies.

The Republican Party finds itself in a weird squeeze on the AI question. Many GOP lawmakers and governors have enthusiastically backed Big Tech’s AI build-out as an economic and national-security imperative. President Trump’s pro-AI infrastructure agenda — from rapid data center approvals to limiting state and local regulatory power — has been presented as a key front in a global competition with China. At the same time, rank-and-file voters in data center host states are increasingly troubled by the real costs — from utility bills to grid strain — that these facilities impose, leaving Republicans vulnerable to being tagged as defenders of distant tech interests at the expense of Main Street.

Democrats, meanwhile, are trying to make sense of this new landscape. Party standard-bearers like Pennsylvania Gov. Josh Shapiro have heralded AI as the future of the economy of their jurisdictions, while also awkwardly defending the backlash. Some progressives have embraced calls to slow or regulate data-center construction, framing it as an issue of worker protection, environmental justice, and local control. Others worry that overt opposition to AI infrastructure will hand Republicans a cudgel to beat them with on innovation and growth.

There’s a bizarre alignment emerging where figures as ideologically opposed as Bernie Sanders and Ron DeSantis find themselves on the same side. In a post on X over the weekend, Sanders said that the key disruptors of the 20th century economy — automobiles, radio, TV, computers, the internet — pale in comparison to the revolution AI will bring about. This week, Sanders is traveling to Stanford, the heart of the Big Tech beast, with US Rep. Ro Khanna (D-Calif.) to call for a moratorium on construction of new AI data centers to “make sure AI works for workers, not just billionaires.” Meanwhile, in December, DeSantis proposed an “AI Bill of Rights” to safeguard Florida residents by addressing the misuse of personal data and allowing local governments to block new data centers. “I really fear that if this is not addressed in an intelligent and proper way, it could set off an age of darkness and deceit,” DeSantis said during a press conference. When a democratic socialist and Florida’s No. 1 Right-wing culture warrior start sounding like joint authors, something real is happening.

If we look a decade ahead, the AI revolution may force a complete rewrite of the old political spectrum of  “Liberal vs. Conservative”, which will collapse into incoherence. In its place, we could see a structural realignment around — to borrow from Marx — who controls the Means of Prediction.

As AI automates the “cognition economy,” the debate over Universal Basic Income (UBI) or a Universal Basic Dividend (UBD) will shift from Silicon Valley pipe dreams to the center of the debate. Governments will be pressured to tax “robot labor” simply to prevent the total collapse of the consumer economy. If people don’t have wages, they can’t buy the products the AI is making. Just as the 19th-century state had to break up the trusts and monopolies of the Gilded Age, 21st-century politics will be a battle between the federal government and a handful of what Sanders calls tech oligarchies. These companies now control the world’s primary compute resources — the digital equivalent of oil or steel. Or we could see the opposite: a command-and-control economy, as the state and tech formally merge and the trend of state capitalism takes firmer hold. 

The danger, I believe, isn’t that AI will produce a single coherent political movement or a technofascist state (so put away your dog-eared copy of 1984). Instead, it may fracture the old coalitions faster than new ones can form, producing a volatile landscape of backlash, regulatory whiplash, and even more opportunistic demagoguery. On the far Left, we see a rise in techno-skepticism and a demand for the abolition of “bullshit jobs” — the idea that we should stop pretending we need 40 hours of work if the AI can do it in four. On the far Right, nationalist protectionism may not align with the needs of the global AI industry. The future, for better or worse, is up for grabs. 

From this vantage point, the AI revolution is the final insult to the American worker, but over the rest of the 2020s, it may also represent the final test for American democracy. The Third Industrial Revolution taught factory towns that the future could arrive without them; the Fourth is teaching the same lesson to the laptop class. And when millions of people are suddenly told, by code, that they are surplus, it won’t be the moths that get the elites. The wolves will find them first.

Ismail’s palace lies in ruins now, its once-ornate filigree wrecked by desert winds and a 1755 earthquake. The remains of the sultan’s royal house have been incorporated into a residential neighborhood. Moroccan life, at the edge of the Sahara Desert, goes on. But the world still contains kings with the vision and resources to imagine defeating the sand and the wind. Today the most ambitious of these is Neom, a series of trillion-dollar “gigaprojects” launched in 2017 by Prince Mohammed bin Salman, the ruler of Saudi Arabia, with the aim of transitioning his country’s economy away from petrodollars toward a more “sustainable” future. 

The original vision included a “ski village”, an artificial marina, a floating port city on the Red Sea, and “The Line”, a planned city 500 meters high, 200 meters wide, and stretching over 100 miles from one coast of Saudi Arabia to the other. Only now Neom is in flux, reflecting a wider world in which visions of the future are being hurriedly rewritten.

As initially envisioned, it was ambitious on a scale that dwarfs even Ismail’s vision — especially The Line. What some might politely call “a bold vision” others would simply describe as “bonkers”; this vertical city, with its mirrored exterior walls, was to house millions of people, have its own airport, be self-sufficient in food, and run on renewable energy. The original vision included a “hidden marina” accessed through a gate in the mirrored city wall, in the apex of which a “chandelier” building was to hang in mid-air. On the levels above this arch, there was to be a football stadium. 

More from this author

Why the Great Reset failed

By Mary Harrington

According to one estimate, building the first 12 sections by 2030 would have meant buying up most of the world’s supply of steel and cement, and bringing in a 40-ft container of materials every eight seconds, 24-hours a day. The early stages of construction involved digging out 100 million cubic meters of soil for the marina, flattening a mountain for an airport runway, and evicting desert tribes (these were dispersed, arrested, and in at least one case killed). 

The cost was staggering; trillions have already been spent. You can see the groundworks from space. Only now it looks like it might never be completed. Over the weekend, the management group admitted that Neom will be downscaled and reimagined, reflecting the Saudi monarchy’s shifting priorities. 

Is this all just a monument to the hollowness of monarchic ambition? Should we perhaps draw out a homily applicable to the proposed new White House ballroom? The FT reports that some of the pilings are already being covered by sand, and yes, it’s tempting to recall Shelley’s meditation on the fleeting quality of even the grandest ruler:

“My name is Ozymandias, King of Kings;
Look on my Works, ye Mighty, and despair!”
Nothing beside remains. Round the decay
Of that colossal Wreck, boundless and bare
The lone and level sands stretch far away.

But this isn’t just about hubris. Rather, the reworked Neom is testimony, in its abandoned groundworks and re-imagined architecture, to the replacement of one vision for the world’s post-industrial future by another. The updated vision is perhaps better suited to absolute monarchs such as Mohammed bin Salman than what it replaces, but also — for the rest of us — far less utopian. 

It’s difficult to find much in the way of visual reporting online, depicting what’s actually there. What you find instead is lavishly visualized CGI renditions of what it will look like: glossy marketing websites, architects’ models, futuristic videos and a ton of “partner content” product-placement material.

Setting aside for a moment how much of this has been realized, or ever will be, in terms of aesthetic and concept, Neom is the Platonic form of an architectural style and overall social order I think of as “Airport Hyperreal”. “Hyperreal”, because the best buildings in this style employ modern construction techniques to play with form, making solid surfaces seem airy and building materials seem less solid than expressions of pure idea. And “Airport” because it’s a style often employed for airports, but also because airports as such are central to the post-industrial, post-national information economy that tends to flourish amid this architectural style. 

Britain has a share of this placeless meta-civilization, but it’s not our strong suit. Peak Britain was the previous age, back when Shelley wrote Ozymandias in 1817. With the 1755 earthquake that wrecked Moulay Ismail’s palace still within living memory, Britain was in her industrial, imperial pomp — an era characterized by frenzied construction of houses, factories, and infrastructure alike. Today, those days are over: it’s now so difficult to get large-scale works agreed in Britain, let alone built, that the 359,866 sheets of consultation paperwork for the proposed Lower Thames Crossing tunnel would, if laid end-to-end, reach almost the whole way from the proposed tunnel site to Calais. As for the “fish disco” at Hinkley Point, and £100m “bat tunnel” of HS2, the less said the better. 

What we have, instead of renewed and working infrastructure, is patches of concentrated overgrowth in the Airport Hyperreal style, perched atop the ruins of imperial Britain. Perhaps the most mind-bending version of this is the City of London, some of whose walls date back to Rome but whose lower-rise buildings are mostly Georgian and Victorian — now dwarfed by extravagant, sinuous glass skyscrapers that seem almost to float in mid-air. The same quality of architectural palimpsest has recently become visible in contemporary photos of rapidly developing Manchester as well, where mid-rise, industrial-era brick huddles before brand-new steel and glass monoliths.

Beyond the scale and choice of materials, what separates Airport Hyperreal from its industrial precursor is its attitude to public space. Where industrialists built open-access parks, museums, galleries and libraries, Airport Hyperreal tends to focus resources on exclusive, gate-kept ones. The paradigmatic example is the airport lounge, where higher status travelers can escape the chaos and anomie of the main concourse for a better-resourced and more genteel atmosphere. The same built-in exclusivity can also be seen in the acrylic swimming pool suspended between two skyscrapers at Embassy Gardens in Vauxhall. This is pure Airport Hyperreal, both in gleefully flouting every common-sense stricture of building design, and also in its flamboyant exclusivity: a leisure resource that ordinary passers-by cannot access, made into an eye-catching design feature.

Neom was clearly envisioned for this kind of society. Not, that is, a completely “open” one of un-gated public space, but something at most semi-exclusive. We might call it airport-lounge urbanism: a country-wide program of building projects, designed for a semi-exclusive, upscale society where manufacturing happens elsewhere, and people divide their time between knowledge work and consumer leisure. A port city; resorts and hotels; a high-density, “sustainable” city with plenty of office space but no factories; a “ski village” comprising real and artificial slopes; a sports stadium half a kilometer into the air; a harbor for cruise ships. 

So how should we understand the fact that Neom has been scaled back? The foremost reason, it has been reported, has been unanticipated budget constraints, due to the softening price of oil. In other words: a development whose stated aim was transitioning the world’s foremost petrostate into the post-oil era has ironically been stymied by its reliance on oil profits. But still more significant is the proposed pivot. Instead of creating a sprawling complex centered around a presumed population of millions of knowledge workers, Neom may be re-imagined as “a hub for data centers”, as part of “Prince Mohammed’s aggressive push for the kingdom to become a leading AI player”. 

Anyone who makes a living from, within, or predicated on the kind of knowledge work that happens in Airport Hyperreal architecture should see this for what it is: a dread omen. It is, to date, the largest-scale physical expression of a general trend, among the world’s movers and shakers, that we might call the “Greater Reset”. 

Suggested Reading

How Saudi bought Britain’s architects

By Malise Ruthven

The “Great Reset”, a program proposed during Covid by the World Economic Foundation, came swiftly to be decried as a conspiracy theory. But on a more charitable reading, it was intended benevolently: as a means of shifting the post-industrial world toward universal, “sustainable”, tech-optimist utopias. It wasn’t popular with the general public, not least because people accurately intuited how much soft authoritarianism and social engineering would be needed for its realization.

But however hubristic its vision, and however disingenuously it was designed to exclude the riff-raff, the “Great Reset” vision expressed architecturally in Airport Hyperreal did at least try to imagine utopias for many humans. By contrast, the Neom rethink shows the contours of the Greater Reset: a future where humans are largely obsolete, except to tend vast datacenters, as ever-greater swathes of knowledge work becomes automated. This future is showing signs of having arrived in Britain: our brief overgrowth of Airport Hyperreal architecture is already threatened by the Greater Replacement, with national net job losses of more than 8% over the past year, due to AI. Meanwhile, even private individuals are embracing AI agents such as “Clawdbot” for tasks that a secretary, administrator, or services employee might once have done, such as booking restaurants or even managing finances.

And yet perhaps Britain has less to worry about than some places. Even if the Greater Reset is grim in the short term, London has been there a long time. The still-visible traces of Roman stonework attest to the city’s ability to endure even tremendous upheaval, with or without skyscrapers and spreadsheets. By contrast, the desert monarch Prince Mohammed bin Salman has evidently decided the international laptop class is no longer a sufficiently promising bet to warrant trillions of infrastructure investment. 

So now Neom is still about housing, of a kind. It’s still futuristic. But it’s not primarily for humans — at least, not very many of them. We might ask the architects: what will become of the people thus replaced? You might as well ask Moulay Ismail to care about the death of a single slave in the building works at Meknes. 

“There must have been about 10 apartments and there were 600 people,” he tells me. “And the situation looked so bad that a number of old people climbed up onto the roof and jumped off.” 

When not chewing the fat with me in the pub, John spends his time speaking in schools for the Holocaust Memorial Trust. Lest we forget. 

Of the remaining 200,000 or so Holocaust survivors, most are in their late eighties, which means that these vanishing witnesses to Nazi evil may all be gone within a decade. That is why, a few years back, John took part in Testimony 360, an innovative project to create a digital record of his experience. Thanks to some sophisticated AI and virtual-reality technology, future students will be able to interact with a virtual John, ask him questions, and learn from his testimony.

More from this author

Tommy Robinson is a cynical Christian

By Giles Fraser

But the use of AI in Holocaust remembrance is becoming increasingly controversial as fake images circulate around the internet. Such is the level of alarm that the International Holocaust Remembrance Alliance has issued a warning to schools, concerned that memes are being used “for political or ideological purposes, including denial and distortion”. Holocaust memory, they say, is under attack and false narratives are on the rise at a time of significant jeopardy for the Jewish community. 

For the time being, those with a keen eye and some knowledge of historical events can still tell the difference between fake images and real ones. There were only a handful of real photos taken within Auschwitz, and so images of children hiding under floorboards or playing violins ought to be treated with great caution. Fake ones are all too obvious, clearly designed to tug at the heartstrings, and often contain small mistakes. But as AI improves such giveaways will be ironed out. Inevitably, it will become harder to tell the difference.  

For some, the fictionalization of the Holocaust — in literature for example — is a useful way of introducing the young to the reality of the horrors of the genocide perpetrated by the Nazis — take The Boy in the Striped Pajamas, for example. It offers an accessible way into a historical reality that is too dreadful for children to absorb in its entirety. But even this can be controversial. As Imre Kertész, another Budapest-born Holocaust survivor, said in Who owns Auschwitz?: fictionalization can mean that “survivors watch helplessly as their only real possessions are done away with: authentic experiences”.

Whatever we conclude on this tricky question, bot-promoted deepfakes are different; it’s not clear that these images are a fiction at all. And so, as fabricated content floods social media, the truth becomes harder to recognize and easier to dismiss. Suddenly, a whole new world of Holocaust suspicion opens up. The fear is that certain forms of Holocaust remembrance could even be conscripted into Holocaust denial. 

In church, we know how technology can complicate memory. Wedding videographers promise to preserve a perfect record of that special day. Yet many couples find that, over time, this vivid recording crowds out any organic memories of the day, leaving only what the videographer saw. Ten years on, wedding couples admit they can’t actually remember those little details that made the day so real to them. Recorded “memories” can easily supplant real ones. 

In handing over responsibility to the videographer, you also relinquish control over how the day is curated and, potentially, remembered. AI imagery can do the same with history: creatively and powerfully interpreting historical events through particular perspectives, until interpretation muscles out the truth. 

Memory is the lifeblood of Jewish togetherness. Jews speak of the deceased as being “of blessed memory”. Passover is an act of collective liturgical memory, remembering the liberation from Egyptian bondage. “Remember the Sabbath day and keep it holy” exhorts the book of Exodus. The Hebrew Bible is full of instructions to remember what God has done. In the UK, we mark Holocaust Memorial Day today, the 81st anniversary of the liberation of Auschwitz. In Israel, remembrance takes place on Yom HaShoah in April or May (27th Nissan, to be precise), the anniversary of the Warsaw Ghetto uprising. The different dates register different emphases: one the liberation of passive Jews by others, the other the heroism of Jews who resisted the Nazi regime. But memory is the common thread. And threats to it are rightly regarded as existential. 

Suggested Reading

History has caught up with British Jews

By Tanya Gold

This is particularly salient today, amid the continued fallout from the October 7 massacre. Even though the Holocaust is the only historical event that is compulsory within the KS3 history curriculum, schools are becoming increasingly nervous about commemorating it. Only this month, a visit by a Jewish MP to a school in Bristol was canceled following protests by local pro-Palestinian groups. And over the past few years, the number of schools marking Holocaust Memorial Day has halved. 

There are some, Simon Jenkins for example, who argue that we should do less remembering and more forgetting. It is an unusual take for a historian. He warns us not to “drink from the rancid well of grievance”, and that “almost all the conflicts in the world are caused by too much remembering: refreshing religious divisions, tribal feuds, border conflicts, humiliations and expulsions”. I take the point but I don’t agree. Memory makes us who and what we are. It creates shared identity. That’s why so many of us fear dementia. To lose our memory is to forget who we are. 

There used to be graffiti on the towpath near our pub. Six million lies, it read. The graffiti has been washed away. But the lie has not. Memory, then, is both a reminder and a warning of what we are capable. Memory is what renews our conviction: never again.

And not just any old bikini deepfakes, even though many AI platforms are capable of creating them. It’s easy to do: you just say to the AI “Put her in a bikini” and the politician, celebrity, or — sometimes — child is suddenly wearing only a bikini. Yes, it’s creepy. But also striking is the way Labour’s ire is laser-focused on Grok, the AI attached to X, formerly Twitter. It’s also striking how, despite pleas from campaigners for years on the platform’s abundance of obscene material, it’s only now that Labour is swinging into outraged action. So grievous is the danger posed by its bikini deepfakes in particular (including of Starmer), that the Government now plans to outlaw them, even as it’s  threatened to ban X from Britain altogether.

This is where the political expediency bit comes in. In Starmer’s shoes I too would be itching for an excuse to ban X. Under Musk’s ownership, the site has made Starmer’s political life a misery, more or less from the moment he entered Downing Street, from stirring the pot during the Southport riots to amplifying the grooming scandal to the point that Labour were forced to respond with the Casey review. And the response to the review, again largely convened on X, forced Starmer to promise the national inquiry his party is now doing its best to avoid delivering.

This is a far cry from pre-Musk Twitter, which was a liberal-dominated hub of respectable media-class opinion, one with a direct line for governments and intelligence services seeking speech-control influence. But since Musk bought the site in 2022, it’s morphed by degrees into something far more pungent and anarchic, where Right-wing voices predominate and many of the former progressive Twitterati have fled to the Bluesky echo-chamber or simply quit posting.

This has not been an improvement in every respect. I don’t miss the sanctimonious tone of old Twitter, but it was more playful. Even so, the rambunctious Right-wing X is politically significant, not least in being the sole large-scale digital exception to the general rule that major institutions are controlled by the Left. As such, the site plays a pivotal role in UK politics. Whether feeding the “posting to policy pipeline”, or drawing attention to ham-fisted efforts at propaganda or speech control, X convenes so large and vocally anti-Starmer a subset of the British Right that we can hardly blame Keith for wishing, like Henry II, that someone (in this case Ofcom) would rid him of this turbulent website.

More from this author

Elon Musk: neo-feudal prince

By Mary Harrington

So is this just about Starmer’s ailing premiership? Yes and no. He’s at the helm of a regime that is, in its entirety, holed below the waterline by the subversive uncontrollability of digital public life. He can hardly be blamed for lashing out at a website whose denizens, and owner, so obviously loathe him. But the proposed ban is best understood as part of a wider war — as was the killing of Thomas à Becket.

Henry II had his Archbishop murdered in 1170 in the course of a dispute over the competing political and spiritual prerogatives of Crown and Church: a contest that obliged both parties to turn to the Pope in Rome for arbitration. And today Starmer finds himself at the mercy of not exactly a religious but certainly a cultural force, one whose extra-territorial leader also has an agenda that doesn’t always (or even very often) align with Starmer’s own. There is of course nothing very holy about Musk, or the trolls of X. But the contest over bikini deepfakes points toward a larger-scale struggle, with cultural, political, and economic dimensions and between overlapping Old and New World jurisdictions.

The field on which this is generally litigated is “free speech”, typically referred to as if it were an abstract moral value. But while there surely is a moral dimension to free speech, in an information age the power to free or control “speech”, and to determine what counts in that definition, is of immense political and economic consequence.

We are, in other words, once again litigating the extra-territorial reach of something intangible, much like Henry II. But in the intervening period Britain also gained her own territorial empire: and then lost it again, at the hands of our own former colony, in a way that also provides lessons for our situation now. In 1941, the British Empire, struggling against the Axis, was obliged to negotiate for aid from its secessionist progeny America. A key dispute within the negotiations was over “imperial preference”. This was a set of tariff arrangements enacted within the British Empire which gave preferential trading rights to polities that were part of the Empire. This so gravely disadvantaged American trade that its dissolution was a cornerstone of mid-century US foreign policy. Britain gave up on imperial preference and secured American aid.

I mention this not to re-litigate long-dead geopolitical quarrels, but to draw a parallel with the broader context of Starmer’s proposed ban of X. This modern argument doesn’t concern physical trade in goods, but the rules under which the digital kind are permitted to circulate. Many of these “goods” — including everything that happens on X — fall broadly under the heading of “speech”. In such a context, who gets to set the rules on “free speech” is obviously every bit as important as the 1941 battle over who gets to determine tariffs across British imperial territories.

As things stand, the digital “imperial preference” favors the polity in which the internet was first developed: the United States.  The question of geography is notoriously vexed when it comes to trade. Even determining who has tax and regulatory jurisdiction over (say) a pile of grain in a warehouse forms the topic of intense negotiations between trade experts. Intangible goods are still more complex, though: when a website can be served in multiple physical locations simultaneously, who gets to regulate it?

Extensive material and political interests rest on how this question is settled. And this is the real geopolitical backbeat to Starmer’s feud with Elon Musk: a backbeat that, in truth, encompasses not just Britain but also the EU to which Starmer clearly wishes Britain still belonged. It’s a feud currently playing out in a series of bitter skirmishes between the White House and the EU, as well as, in microcosm, between Musk and Starmer.

Recently, for example, US Secretary of State Marco Rubio announced sanctions against foreign individuals or institutions involved in censoring US digital properties: a clear shot across the bows of European politicians who believe they have a right to censor US websites and fine the companies  where these are served on EU territories. In the pandemonium that erupted following Rubio’s announcement, pro-EU supporters pointed out that no such challenges had been issued against Russia or China, despite these polities firewalling large portions of the American internet in its entirety.

But this is precisely the rub. Do the European states want to be as detached from the US as Russia and China, or not? There is no freedom of speech battle between these polities and the US, because neither Russia nor China tries to control what’s published on American platforms (at least, not openly). They just block them, and build their own platforms. By contrast, European states seem happy for America to supply its digital platforms, even as European politicians demand a say in how they are regulated. It’s as though 20th-century Canada was demanding, on the basis of its extensive wood and paper exports to Britain, a veto over everything published in London.

Until relatively recently, this relation of center and periphery was broadly mutually beneficial — not least in that it enabled American politicians to route round inconvenient constitutional clauses at home, in regulating the internet. Under Biden, the State Department encouraged and fostered European online speech controls, as a means of developing censorship tools both unconstitutional in the USA but nonetheless applicable domestically where desired, and aligned with an overall policy of digital speech control. I’m old enough to remember when something being illegal somewhere in Europe could result in its being deleted altogether from X even if the poster was in the USA. In effect, the Biden regime used Europe as a way of offshoring politically sensitive speech controls, somewhat like Guantanamo Bay offshored the torture of suspected jihadists.

More from this author

The tragedy of Keir Starmer’s Substack

By Mary Harrington

Now, though, priorities have changed — much to the bewilderment of Europe, which now has a well-developed censorship architecture but has discovered its American sponsor suddenly in a different mood. Even so, they have pressed on: two heavy-hitting US websites, X and Cloudflare, have already been threatened with fines under, respectively, the EU’s Digital Services Act and Italy’s “Piracy Shield”. The US-based forum 4chan has also been slapped with fines under Britain’s Digital Services Act. All have responded, essentially, “GFY”. Starmer’s threat to Musk on the deepfake image-making falls in the same category. It is an attempt to exert sovereignty in a space that’s both nominally open, but also in practice wholly dominated by American “imperial preference” — in the context of an American presidency taking a radical new direction.

Seen thus, the actual contest becomes clearer. Is Europe an American imperial subject, or a rival imperium? If it’s part of the empire, then obviously it doesn’t get to regulate its hegemon. If Europe is an imperial rival, it can hardly expect to demand its opponents supply services even as Europe dictates the terms. This is the real subject of the ongoing diplomatic wrestling-match over “free speech”, censorship, and regulatory frameworks such as the EU’s Digital Services Act and Britain’s Online Safety Act. If Britain, or any other European entity, really wants information sovereignty outside the American imperium, the only route there is the one now threatened by Starmer: banning American digital properties entirely from our territories, the way Russia and China do.

Does Starmer really want to go there? Does any EU state? This is perhaps uniquely difficult for Britain to answer. Geographically and economically speaking, Britain is closer to Europe, which accounts for around 41% of UK exports in 2024, compared to 18% for the US. But linguistically and culturally we’re closer to the USA. To take just one of countless examples, the BBC pays scant attention to EU politics, but Trump only has to raise an eyebrow and it’s front-page news.

In this context, can we really imagine an Atlantic firewall? I doubt it. Back in 1170, ridding himself of Thomas à Becket backfired badly on King Henry; I suspect that ridding himself of this turbulent X would prove similarly unpopular with the British media and political class. And yet our cultural, linguistic and digital enmeshment with an increasingly unpredictable US regime has introduced an acute new set of tensions. It is difficult to see how they can be resolved peacefully. But let us hope they can be. Having seen one bikini’d AI Starmer, I have no desire to see any more.

There are reasons to wonder if these poles, and much else in the Rockefeller Wing, should be on Fifth Avenue. But here they are. And wandering through the halls recently, the late afternoon sun slanting through the window-walls, I realized how badly I had missed them for the last four years — while all of this was closed for renovations. When I first entered the southmost part of the room, where the bis poles used to stand, and they weren’t there, I panicked for a moment. Had they been… repatriated? I accepted that if so, it was probably quite correct. But I felt depressed. I sidled up to a guard in big shoes, and asked if there wasn’t some stuff missing in this room. He said no, that stuff was now next door.

Sure enough. Standing there beneath these human trees, strangely relieved, I was suddenly struck by the sense that the room in which I was standing — each elegantly balanced case and each careful label — delivered a precisely aimed punch in the nose for our moment, a wake-up slap for this bad dream.

After all, we in the US — as well as folks in the UK and Europe — are in the throes of intense conflict concerning “diversity”. In recent years, a longstanding and broadly operational national consensus around affirmative action policies (aimed at addressing legacies of racial and gender discrimination) has cracked rapidly into hostile camps: on the one hand, defenders of DEI policies and progressive identity politics; on the other, defenders of the “Enlightenment ideal” of free and equal persons judged on their substantive merits, without reference to sex or skin color. Each side accuses the other of bad faith, in addition to bad ideas. If it’s diversity you want, say MAGA pundits, how about we get some Republicans into your philosophy department? If you hate non-white people, counter the Brooklyn socialists, why not just say so, instead of pretending that you care about free speech?

Meanwhile, the Asmat bis poles look on, and they must be laughing hard in their woody cores: “Diversity?” one can hear them thinking, “How dare these cretinous buffoons babble on about diversity!”

Suggested reading

Big Tech has stolen our children

By Matt Feeney

After all, those looming funeral pylons, together with all the other majestic Melanesian artifacts around them, which hail from hundreds of totally diverse cultures, can see clearly the truth to which we are mostly blind: our world, the world of modern America and much of what lies beyond these shores, is, at this point, almost perfectly monolithic. Big picture, we all share one simple cosmology. Call it the “moneyworld”.

I know what you’re thinking: “That can’t be right! Look around! We are Christians and Jews, atheists and futurists, men and women and everything in between, black people from the South and recent immigrants from Asia. We are all so distinctive. And we believe a million different things. How dare you call us all the denizens of a single cosmology — and a counting-house ‘moneyworld’ at that!”

Almost. But not really. What we are, mostly, of course, is moderns, and what that means is that we navigate a world mapped in money values. A world shaped by ever-more-powerful programs for aggregating and augmenting “wealth” — wealth measured in dollars and cents. The masters of our universe are the financiers who look ceaselessly for ways to make their capital return new and bigger profits. We follow them, and pay tribute, when we can, with our little mutual funds. And we, too, as consumers and life-long market players, try to figure out how to win in the ledger book of existence: to buy the stuff we want for prices we can pay; to get ours at the auction block where we sell our wares, our time, our eyes, our minds. Indeed, we play the market with our whole beings, since we do everything we can to be rational self-maximizers in the continuous competition of the workforce, which demands that we conceive of ourselves as human capital, and “invest” in ourselves, through education and training, to optimize the ROI of our lives.

So what matters to the moneyworld? Money matters. What is “meaning” in the moneyworld? A billion dollars is meaning — everyone gets that. Are there other kinds of meaning? Sure, maybe. Like “happiness”, perhaps? True, we know it is a thing. But the good news is that economists can make you a model for it — denominated in dollars, naturally (they have no other units). What their models cannot capture is squishy at best. It might be real stuff, but it is so hard to quantify. Better pick a major in college that pays real dividends. In cash. You can puzzle about happiness after you make your first million.

Let me be clear about my view here: money is fine. Yes, yes, it is indeed a powerful technology for coordinating human behaviors, and markets can do good things at scale — allocate resources, drive innovation — in ways that are unmatched by any other social system or set of beliefs. I get it.

But the moneyworld is a world organized by price, a world of pure and continuous calculation. It is a world that only a machine could love — and the machines do love it. They run the numbers and keep the accounts. They have the data, and they use it — to monetize us, actually. Indeed, it is increasingly their world, and we only live in it — with many of us feeling more and more alienated and angry and unsure of what comes next.

A turn in the Oceania gallery of the Met is a chance to be reminded of how strange our machinic moneyworld really is. It has engulfed the planet as a ubiquitous system of meaning and order — but it has no soul whatsoever. No imagination. It knows, in the end, nothing about the stuff that makes us what we are as human animals: the dreams, the nightmares, our unique ability to find each other and tell stories powerful enough to make the earth a home and turn perfect strangers into kin — all the mad and magnificent ways we have worked to make sense of the experience of being alive on this planet. What makes us human, ultimately, is our ability to make meaning, not money. And we have done that in millions of ways, across millennia.

In New Guinea alone, a single island, there are some 800 different language communities. In some regions a mere 50 miles separates groups who are farther apart in linguistic and cultural traditions than China and France. How different are these worlds? Very. Look over here, at this “yam mask” from the Abelam people of the Middle Sepik River, over the ridge from the Asmat. It was placed on giant, man-sized tubers in rituals of fertility and initiation. It sits in a case, here, as a small and mysterious fragment from a very different cosmology of meaning and practice. Little of that world survives, of course: the grandsons of the makers of that mask probably work for the moneyworld now, logging in Malaysia.

Will a stack of greenbacks someday sit in a museum case on Mars, as artifacts of our globe-spanning but heedless and apparently self-destructive worldview? Who can say. But I like to imagine what the curators would write on the label.

For now, we have our work cut out for us. It is time to dig in and think hard about the different ways our ancestors have made meaning. Because what it is to be a human being is an urgent question to which every one of us had better have an answer. The rise of super-powerful human-emulator AI systems — which are both the flower of the moneyworld and its killer app — has made the question of our actual humanity the blazing question of our time.

We have been and believed so many things. That diversity, now, is a powerful inspiration as we face the monolithic moneyworld we have made. How can we arm ourselves against dehumanization in all its brutal forms? I don’t know. But I do know that there is no whitepaper that is going to save us. And that our museums, most of them, remain open. Some of the thinking we need is going to happen in their halls. It must.

For all the pioneering smartphone refuseniks and buzzy new alternatives, the data tells a less optimistic story. In Britain, Nokia brick sales jumped from 400,000 to 450,000 in 2024, but 68 million people have smartphones. Screentime restriction apps and all sorts of lobotomized dumbphones are not actually flooding the market. The most hyped post-screen pioneers — the Humane AI Pin and Rabbit — were honorable but catastrophic failures.

The situation is critical. The scrolletariat is swiping itself into oblivion. While distinctions between productivity and entertainment make measurement tricky, the average human spends roughly 4.9 hours on mobile apps daily. For the young, it’s even worse. American teenagers average a staggering 8 hours and 39 minutes of daily screen time for entertainment, while British 18-24 year olds clock in at nearly 6 hours and a half hours. The scale is vast, and with the rise of hyper-personalized AI slop machines, the quality is degrading further. Add to that the wider outsourcing of human cognition to AI, we risk becoming hollowed out fleshy middleware lost in an endless mind maze.

In the face of this smartphone-AI-dopamine axis of evil, various schools of neo-Luddite resistance are emerging. Like the loom-smashing weavers of the 19th century, today’s resistance is united by a distaste for the machine. However, they are divided on the best course of resistance — many want to run for the hills, others want to destroy their phones, and some believe they can tame the effects. All are doomed to fail.

To those with a more romantic vision of resistance, the digital world is only avoided by escaping into a prelapsarian Arcadia of nature, craft, literature, and “dumb” technology. They range from the Bopeas trading screen time for soil and wild swimming and the hip Gen Z resistance groups such as the gnome-hatted Lamp Club and Breaking the G(loom) all the way to noted anti-machine rager Paul Kingsnorth and The Times’ James Marriott wielding his Qin F21 Pro. They all instinctively understand the sickness, but their cure is fundamentally unscalable. Unchaining oneself means battling the network effect, the mechanism that locks users in by punishing non-participation with isolation. Escape therefore requires the social capital to survive being cut off from the herd — a luxury most cannot afford.

Equally romantic, but much less chill are the more violent neo-Luddites. These are the spiritual children of the late Ted Kaczynski, the notorious Unabomber who resisted industrial society through manifestos and mail bombs. They share a quasi-millenarian view of techno-industrial society, seeing the destruction of the mega-machine as essential to the preservation of nature and humanity, and are willing to break some legs to get there. These neo-Luddite extremists include the Greek anarchists Conspiracy of Fire Nuclei, the mystical Latin American terrorist networks of Individualists Tending to the Wild, and France’s emerging Anti-Tech Resistance. While they may be true believers, tying salvation to the destruction of techno-capitalist civilization is not a winning formula with the less insane.

More pragmatic is the network of politicians, policy wonks, academics, and journalists pursuing reforms through legislation. On the Left, Shoshana Zuboff, American author of The Age of Surveillance Capitalism, and British journalist Carole Cadwalladr want to tame the techno-beast through privacy, anti-trust, and online safety bills. On the Right, there’s a mirror universe of pro-family tech critics, from US Senator Josh Hawley through to Tory MP Miriam Cates. While legislation is crucial, attempts are glacially slow and critics are fundamentally divided on the true source of evil. Are screens the enemy? Or surveillance? Hate speech? Big Tech? All forms of social media? Dopamine culture? Governments must play a role, but for now we are in a race against time. Endless hubs, quangos, and inquiries will not outrun the shapeshifting onslaught of innovation. Even if they do eventually succeed, citizens will bypass them.

Further, governments seeking to address the neo-Luddites’ concerns face a paralyzing paradox. While they may want to halt mass cognitive atrophy, other crises take priority. Faced with anemic growth, declining living standards, geopolitical instability, and climate pressure, government ambitions are increasingly beholden to Silicon Valley. Leaders see AI as the silver bullet in the hunt for productivity, security, and sustainability. Yet, the very titans building these frontier technologies are the same ones powering the brain rot. This Faustian pact is laid bare by the British Government’s new partnership with Alphabet to harness AI for clean energy research and public services. If Google is crucial to the state, tackling YouTube addiction is politically impossible.

Thus, we are at an impasse. If we cannot yet regulate through force or law, what is to be done? Perhaps there is a way to become masters of our domain and usher in a grand transformation of social norms.

We have been here before. The Luddites failed to stop the Industrial Revolution quashed by an alliance of government and merchants comparable to today. Industrialization continued and unleashed a different blight: cheap hard booze. From schools and asylums to factories and prisons, everyone was getting lit. The 18th-century Gin Craze, immortalized in Hogarth’s Beer Street and Gin Lane diptych, was unstoppable. Despite the Gin Acts, the state was captive to the alcohol lobby and tax receipts. Those in power, it was often said, wanted workers too drunk to organize politically, but not too drunk to operate machines effectively in their factories.

In the 19th century, the Temperance movement did what law could not. In September 1832, in a converted cock-fighting pit in Preston, Lancashire, a cheesemonger named Joseph Livesey and six working-class comrades — the Seven Men of Preston — signed a pledge of total abstinence. They believed a drunken proletariat could never be liberated and so-called “Giant Alcohol” was the ultimate tool of subjugation.

Their genius was to reframe the issue. They advanced the debate beyond religious sin to a campaign for wealth, health, and inter-generational morality. Temperance activists argued that alcohol stole a father’s wages, wrecked the worker’s liver, and shattered the peace of the home. As Benjamin Disraeli’s physician Joseph Kidd wrote in 1879, the goal was a virtuous citizen “accustomed to self-restraint and moderation”, a state he called “one of the chief delights and aspirations… of a true man”. The mission was not just to change laws, but to change the soul.

More from this author

The curse of the scrolletariat

By Louis Elton

Crucially, campaigners invoked the innocence of the young. Founded in 1847, a society called the Band of Hope inspired children to sign a pledge of sobriety. It aimed to inoculate children against alcohol through a cultural program including its iconic magic lantern shows, using cutting-edge technology to project vivid, terrifying images of the drunkard’s decline. British alcohol consumption finally peaked in 1876. By 1897, the Band of Hope had enrolled 3 million children. By 1931, a Royal Commission officially declared that “drunkenness has gone out of fashion”.

What can we learn from the Temperance movement? In his 2025 Reith Lecture, “Fighting for Humanity in the Age of the Machine”, Dutch historian Rutger Bregman argues that the Temperance movement is the perfect blueprint for our time. He warns that “today, we face a new addiction industry, not of wine and whiskey, but of apps and algorithms, [where ]many of Stanford’s brightest minds are building a single great moloch, an attention-hijacking machine that devours our focus”. Bregman proposes using the state to break this machine, advocating for dopamine taxes on ad revenue similar to alcohol taxes. Yet, given our governments’ increased reliance on AI companies, these changes seem unlikely to happen soon enough. If we cannot yet liberate the soul through regulation, then perhaps changing norms is the method.

This cultural rewiring begins with a return to collective accountability. Just as with the Seven Men of Preston and the Band of Hope, this begins with a pledge to personal and collective accountability. There are several leading candidates for a modern neo-Temperance movement such as America’s Wait Until 8th, which encourages parents to delay giving their child a smartphone until the end of 8th grade, and Britain’s Smartphone Free Childhood (SFC), pledging to restrict smartphones until 14 and social media until 16. The latter is particularly impressive: having only been founded in February 2023, it has already amassed over 300,000 parents — including Kate Winslet and Paloma Faith.

Will Orr-Ewing, who co-founded SFC’s Smartphone Free Schools campaign, explains that this is fundamentally about re-establishing lost boundaries. He describes the pact not merely as a petition, but as a “covenant that you sign with other parents” to hold the line collectively. The goal is to revive “the Victorian idea that children are a separate category”, establishing childhood as a “sacred time that is given special prominence” and protected from the encroachment of adult vices.

Suggested reading

Techno-politics is fuelling public anger

By Jacob Howland

While we chew through the messy battle to regulate addictive algorithms, neo-Temperance activists are attempting to ensure existing laws are interpreted to factor in novel technologies. Orr-Ewing is already pursuing this, recently launching a Judicial Review to force schools to apply existing safeguarding statutes. He argues that schools have failed to understand that their safeguarding obligations apply to smartphones just as they do to a classroom ceiling full of asbestos. “Whether it’s a beheading video or high-definition porn,” Orr-Ewing argues, “nine-year-olds should not be watching this shit. I think we’ve still got enough moral norms to have consensus on that.”

Beyond the classroom and the courtroom, success requires rediscovering the lost art of judgment. The polite relativism of contemporary liberal culture has trained us to avoid criticizing others. Yet, if we stand by as the minds of tomorrow are obliterated, politeness is complicity. It may feel rude and fogeyish but we must scold — and be scolded. Take inspiration from Chile’s outgoing Millennial President Gabriel Boric who recently paused a speech to phoneshame 11 audience members for staring at their screens before showing off his own dumbphone. We can all harness the spirit of Boric. Be judgy. Call your friend a swipewanker. Ban phones at dinner. Smartphone zombies must be rendered socially toxic. Norms can only be transformed through a collective effort to hold each other to account.

Yet, stigma alone cannot remake a people, one needs stories and symbols woven into culture. The original Temperance movement understood this well. In art, following Hogarth’s lead, George Cruikshank’s wildly popular The Bottle (1847) series depicted a respectable British family’s descent into poverty and madness due to drink. In literature, Anne Brontë’s The Tenant of Wildfell Hall (1848) shattered the Victorian illusion of the merry drunk, exposing the brutal horror of addiction. In mass entertainment, the Band of Hope’s magic lanterns inspired millions. To succeed today, the neo-Temperance movement must find and celebrate its very own Hogarths, Cruikshanks, and Brontës and propagate their works in the physical world. Who will paint the Beer Street and Gin Lane of the smartphone era?

Finally, a neo-Temperance movement will fail without a positive vision for what humans actually ought to be beyond mere pleasure–satisfying machines. Who is the post-smartphone human? What should they do instead of staring and swiping? Gabriella Nguyen, the 24-year-old founder of Appstinence, an anti-social media campaign, argues we do not need to solve this immediately. To avoid simply trading screen time for increased productivity is reductive, so first we must return to “ground zero” to work out how to live. Then, through Appstinence, we can reclaim these “luxuries of time, practice, social support, and a clear mind” and begin to “no longer see these meandering, unbound exchanges as lavish expenses of our time and productivity, but the regular course of events”. It is here, Nguyen claims, we can reflect and forge the post-smartphone human.

At this rate, billions will spend 25 years of their lives on their smartphone. Yet, mass sloppification is ultimately a choice. Neither government nor Big Tech is going to change things anytime soon. The neo-Luddites will not smash the data centers. The blueprint of the Temperance movement reveals that salvation will not come from above — it must be reclaimed from below. The solution is not a single law or tax, but a collective recovery of cognitive sovereignty. It’s time for neo-Temperance.